Make changes to the Mac WebKit2 Skipped file trigger Windows builds
[WebKit-https.git] / WebKitTools / ChangeLog
1 2010-11-17  Adam Roben  <aroben@apple.com>
2
3         Make changes to the Mac WebKit2 Skipped file trigger Windows builds
4
5         We use that file on Windows!
6
7         Fixes <http://webkit.org/b/49643> Changes to
8         LayoutTests/platform/mac-wk2 should trigger a Windows build, but don't
9
10         Reviewed by Steve Falkenburg.
11
12         * Scripts/webkitpy/common/config/build.py:
13         (_should_file_trigger_build): Added the mac-wk2 directory and make it
14         trigger builds on SnowLeopard and Windows.
15
16         * Scripts/webkitpy/common/config/build_unittest.py:
17         (ShouldBuildTest.test_should_build): Added a test.
18
19 2010-11-17  Satish Sampath  <satish@chromium.org>
20
21         Reviewed by Jeremy Orlow.
22
23         Clear the speech input mock explicitly before each test.
24         https://bugs.webkit.org/show_bug.cgi?id=49660
25
26         * DumpRenderTree/chromium/LayoutTestController.cpp:
27         (LayoutTestController::setMockSpeechInputResult):
28         * DumpRenderTree/chromium/LayoutTestController.h:
29         * DumpRenderTree/chromium/WebViewHost.cpp:
30         (WebViewHost::speechInputController):
31         (WebViewHost::reset): Invoke speech input mock's clearResults method.
32         * DumpRenderTree/chromium/WebViewHost.h:
33         (WebViewHost::speechInputControllerMock): Moved mock from LayoutTestController.
34
35 2010-11-17  Alexander Pavlov  <apavlov@chromium.org>
36
37         Unreviewed, specify IRC nickname in committers.py.
38
39         * Scripts/webkitpy/common/config/committers.py:
40
41 2010-11-17  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
42
43         Reviewed by Tony Chang.
44
45         [NRWT] Make http locking similar to perl implementation
46         https://bugs.webkit.org/show_bug.cgi?id=49187
47
48         * Scripts/webkitpy/common/system/file_lock.py: Added.
49         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
50         * Scripts/webkitpy/layout_tests/port/http_lock.py:
51
52 2010-11-17  MORITA Hajime  <morrita@google.com>
53
54         Reviewed by Kent Tamura.
55
56         [Chromium][DRT] EventSender.contextClick() should aware spellchecking
57         https://bugs.webkit.org/show_bug.cgi?id=49366
58
59         - EvenSender: Checked WebContextMenuData.misspelledWord and added extra context menu entries
60           according to the spellchecker suggestion
61         - MockSpellCheck: add fillSuggestionList to provide fake suggestions.
62
63         test_expectations.txt will be changed after this change is ported to test_shell.
64         
65         * DumpRenderTree/chromium/EventSender.cpp:
66         (makeMenuItemStringsFor):
67         (EventSender::contextClick):
68         * DumpRenderTree/chromium/MockSpellCheck.cpp:
69         (MockSpellCheck::fillSuggestionList):
70         (MockSpellCheck::initializeIfNeeded):
71         * DumpRenderTree/chromium/MockSpellCheck.h:
72         * DumpRenderTree/chromium/WebViewHost.cpp:
73         (WebViewHost::mockSpellCheck):
74         * DumpRenderTree/chromium/WebViewHost.h:
75
76 2010-11-16  Dave Hyatt  <hyatt@apple.com>
77
78         Reviewed by Sam Weinig.
79
80         Make sure the pref for minimum font size is just explicitly set to 0 every time,
81         since the bots have 1 stuck in their plist otherwise.
82
83         * DumpRenderTree/mac/DumpRenderTree.mm:
84         (resetDefaultsToConsistentValues):
85
86 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
87
88         Reviewed by Tony Chang.
89
90         Rebaseline server: compute diffs client-side
91         https://bugs.webkit.org/show_bug.cgi?id=49640
92         
93         The image diff output from the DRT is pretty bad on some ports (at
94         least the Mac one), so it's better to compute diffs on the client by
95         using <canvas>.
96
97         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
98         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
99         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
100
101 2010-11-16  Mihai Parparita  <mihaip@chromium.org>
102
103         Reviewed by Tony Chang.
104
105         Rebaseline server: display test results
106         https://bugs.webkit.org/show_bug.cgi?id=49626
107         
108         Adds basic result display to the rebaseline server. On the Python side
109         this involves:
110         - Parsing the unexpected_results.json into a dictionary.
111         - Serving it as JSON under /results.json.
112         (the JSON -> dict -> JSON transform isn't strictly necessary right now,
113         but I'll need to have access to the parsed results on the Python side
114         for follow-up changes).
115         
116         On the web UI side this adds:
117         - Markup for display image and text results (expected, actual, diff),
118           and JS for populating it.
119         - Markup for breaking down test results by failure type and directory,
120           and JS for populating it.
121
122         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html:
123         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css:
124         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js:
125         * Scripts/webkitpy/tool/commands/data/rebaselineserver/util.js: Added.
126         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
127
128 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
129
130         Reviewed by Ojan Vafai.
131
132         See Chromium issue http://codereview.chromium.org/5133001/ - we
133         are modifying the build bots to pass in the builder name with
134         the "GPU" string appended instead of appending it in the code.
135
136         https://bugs.webkit.org/show_bug.cgi?id=49636
137
138         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
139
140 2010-11-16  Steve Falkenburg  <sfalken@apple.com>
141
142         Reviewed by Adam Roben.
143
144         Disable LTCG for Windows Release builds. Add new Release_LTCG configuration.
145         https://bugs.webkit.org/show_bug.cgi?id=49632
146
147         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
148         * DumpRenderTree/win/DumpRenderTree.vcproj:
149         * DumpRenderTree/win/ImageDiff.vcproj:
150         * MiniBrowser/MiniBrowser.vcproj:
151         * TestWebKitAPI/win/TestWebKitAPI.vcproj:
152         * WebKitLauncherWin/WebKitLauncherWin.vcproj:
153         * WebKitTestRunner/win/InjectedBundle.vcproj:
154         * WebKitTestRunner/win/WebKitTestRunner.vcproj:
155         * WinLauncher/WinLauncher.vcproj:
156         * record-memory-win/record-memory-win.vcproj:
157
158 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
159
160         Reviewed by Ojan Vafai.
161
162         new-run-webkit-tests: rename TestInfo to TestInput, move image hash to work thread
163
164         Rename the TestInfo class to TestInput to be clearer about its
165         function, and move the checksum-reading code into dump_render_tree_thread
166         to avoid cross-thread access.
167
168         https://bugs.webkit.org/show_bug.cgi?id=49573
169
170         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
171         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
172         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
173
174 2010-11-16  Dave Hyatt  <hyatt@apple.com>
175
176         Fix bustage. Remove the minimum font size pref setting in DumpRenderTree.
177
178         * DumpRenderTree/mac/DumpRenderTree.mm:
179         (resetDefaultsToConsistentValues):
180
181 2010-11-16  Stephanie Lewis  <slewis@apple.com>
182
183         Reviewed by Geoff Garen.
184
185         <rdar://problem/8624267> Leak creating offscreen webview running fast/dom tests
186
187         Use shared DeviceOrientationProviderMock. The old allocation method created a new WebDeviceOrientationProviderMock 
188         for every WebView, and the WebKit API doesn't support that ownership model.
189
190         I also fixed a leak in setMockDeviceOrientation by adding a missing release.
191
192         * DumpRenderTree/mac/DumpRenderTree.mm:
193         (createWebViewAndOffscreenWindow):
194         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
195         (LayoutTestController::setMockDeviceOrientation):
196
197 2010-11-16  Dirk Pranke  <dpranke@chromium.org>
198
199         Reviewed by Ojan Vafai.
200
201         new-run-webkit-tests: log the process id in --verbose mode
202
203         As part of the switch to multi-process mode, we should log the
204         process id when running w/ --verbose.
205
206         https://bugs.webkit.org/show_bug.cgi?id=49571
207
208         * Scripts/webkitpy/layout_tests/layout_package/printing.py:
209
210 2010-11-16  Adam Roben  <aroben@apple.com>
211
212         Only trigger Windows builds when files that we actually use on Windows
213         are changed
214
215         A new Scheduler subclass, PlatformSpecificScheduler, has been added.
216         It uses the new webkitpy.common.config.build module to determine
217         whether a particular change should trigger a build on a particular
218         platform. The Windows builders have been switched to use a
219         PlatformSpecificScheduler.
220
221         The logic to determine whether or not a particular change should
222         trigger a build on a given platform has only been implemented/tested
223         for Windows. I tried to make it easy to add more platforms in the
224         future, but I don't have enough familiarity with all platforms to be
225         able to implement it for them.
226
227         Fixes <http://webkit.org/b/49407> Windows builders kick off builds for
228         lots irrelevant changes (e.g., rebaselining Chromium test results)
229
230         Reviewed by Eric Seidel.
231
232         * BuildSlaveSupport/build.webkit.org-config/config.json: Use a
233         PlatformSpecificScheduler for the Windows builders.
234
235         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
236         (PlatformSpecificScheduler.__init__): Added. Stores our platform, sets
237         up a ChangeFilter that filters to our branch and will call through to
238         our filter method, and calls up to our base class.
239         (PlatformSpecificScheduler.filter): Calls through to
240         build.should_build to find out whether we should trigger a build for
241         this change.
242
243         * Scripts/webkitpy/common/net/build.py: Added.
244         (_should_file_trigger_build): Uses a set of directories and regexp
245         patterns to determine whether the given file should trigger a build on
246         the given platform. As mentioned earlier, this has only been
247         implemented for Windows, though I did try to make some guesses about
248         other platforms.
249         (should_build): Returns true if any of the files should trigger a
250         build on the given platform.
251
252         * Scripts/webkitpy/common/net/build_unittest.py: Added.
253         (ShouldBuildTest.test_should_build): Does some basic testing to make
254         sure we're triggering builds for the right files. It only tests
255         Windows for now, though I tried to make some guesses about other
256         platforms.
257
258 2010-11-16  Simon Hausmann  <simon.hausmann@nokia.com>
259
260         Reviewed by Andreas Kling.
261
262         [Qt] Remove synchronous QWebPage::checkPermissions signal
263         https://bugs.webkit.org/show_bug.cgi?id=46810
264
265         * DumpRenderTree/qt/DumpRenderTreeQt.cpp:
266         (WebCore::WebPage::WebPage):
267         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
268         (LayoutTestController::grantDesktopNotificationPermission): When granting
269         permission, grant it directly on the QWebPage/Frame, that will remember it.
270         * QtTestBrowser/webpage.cpp:
271         (WebPage::WebPage):
272         * QtTestBrowser/webpage.h:
273
274
275 2010-11-16  Andras Becsi  <abecsi@inf.u-szeged.hu>
276
277         Reviewed by Csaba Osztrogonác.
278
279         [Qt] Add isPrinting and setPrinting methods to DRT's LayoutTestController.
280
281         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
282         (LayoutTestController::reset):
283         * DumpRenderTree/qt/LayoutTestControllerQt.h:
284         (LayoutTestController::isPrinting):
285         (LayoutTestController::setPrinting):
286
287 2010-11-16  Balazs Kelemen  <kbalazs@webkit.org>
288
289         Reviewed by Andreas Kling.
290
291         [Qt][WK2] Add a way to turn on tiled backing store in MiniBrowser
292         https://bugs.webkit.org/show_bug.cgi?id=49587
293
294         * MiniBrowser/qt/BrowserView.cpp:
295         (BrowserView::BrowserView): Added BackingStoreType
296         argument to ctor. Propagating it to the QGraphicsWKView.
297         * MiniBrowser/qt/BrowserView.h:
298         * MiniBrowser/qt/BrowserWindow.cpp:
299         (BrowserWindow::BrowserWindow):
300         (BrowserWindow::newWindow): Create the new BrowserWindow
301         with the same BackingStoreType that this instance has.
302         * MiniBrowser/qt/BrowserWindow.h: Added BackingStoreType member.
303         * MiniBrowser/qt/main.cpp:
304         (main): Use tiled backing store if got -tiled command line argument.
305         Avoid redundant copy of command line args.
306
307 2010-11-16  John Knottenbelt  <jknotten@chromium.org>
308
309         Reviewed by Jeremy Orlow.
310
311         Move DeviceOrientationClientMock from LayoutTestController to WebViewHost.
312         https://bugs.webkit.org/show_bug.cgi?id=48506
313
314         This ensures that there is one mock per page WebView / WebViewHost. This is a
315         design constraint of the DeviceOrientationClientMock because the m_controller
316         field can only store a pointer to a single instance of the corresponding
317         DeviceOrientationController.
318
319         Test: fast/dom/DeviceOrientation/no-page-cache.html
320
321         * DumpRenderTree/chromium/LayoutTestController.cpp:
322         (LayoutTestController::setMockDeviceOrientation):
323         (LayoutTestController::deviceOrientationClientMock):
324         * DumpRenderTree/chromium/LayoutTestController.h:
325         * DumpRenderTree/chromium/WebViewHost.cpp:
326         (WebViewHost::deviceOrientationClientMock):
327         (WebViewHost::deviceOrientationClient):
328         (WebViewHost::reset):
329         * DumpRenderTree/chromium/WebViewHost.h:
330
331 2010-11-15  Andras Becsi  <abecsi@webkit.org>
332
333         Reviewed by Andreas Kling.
334
335         [Qt][WK2] Avoid polling in WebKitTestRunner.
336         https://bugs.webkit.org/show_bug.cgi?id=49542
337
338         Make the WK2 testing session about 10-15% faster and decrease random flakiness
339         resulting from timing skew by removing the millisecond polling from TestControllerQt.
340
341         * WebKitTestRunner/TestController.cpp:
342         (WTR::TestController::didFinishLoadForFrame):
343         * WebKitTestRunner/TestController.h:
344         * WebKitTestRunner/TestInvocation.cpp:
345         (WTR::TestInvocation::didReceiveMessageFromInjectedBundle):
346         * WebKitTestRunner/mac/TestControllerMac.mm:
347         (WTR::TestController::notifyDone): Add method stub.
348         * WebKitTestRunner/qt/TestControllerQt.cpp:
349         (WTR::TestControllerRunLoop::instance):
350         (WTR::TestControllerRunLoop::start):
351         (WTR::TestControllerRunLoop::stop):
352         (WTR::TestControllerRunLoop::TestControllerRunLoop):
353         (WTR::TestControllerRunLoop::timerEvent):
354         (WTR::TestController::notifyDone):
355         (WTR::TestController::platformRunUntil):
356         * WebKitTestRunner/win/TestControllerWin.cpp:
357         (WTR::TestController::notifyDone): Add method stub.
358
359 2010-11-15  Darin Adler  <darin@apple.com>
360
361         * Scripts/webkitpy/common/net/bugzilla: Added property svn:ignore.
362
363 2010-11-15  Benjamin Poulain  <benjamin.poulain@nokia.com>
364
365         Reviewed by Andreas Kling.
366
367         [Qt] [WK2] Add Ctrl+L shortcut to Qt Minibrowser
368         https://bugs.webkit.org/show_bug.cgi?id=49544
369
370         Add the shortcut to BrowserWindow.
371
372         * MiniBrowser/qt/BrowserWindow.cpp:
373         (BrowserWindow::BrowserWindow):
374         (BrowserWindow::openLocation):
375         * MiniBrowser/qt/BrowserWindow.h:
376
377 2010-11-14  David Levin  <levin@chromium.org>
378
379         Reviewed by Daniel Bates.
380
381         check-webkit-style should detect PassRefPtr usage in functions.
382         https://bugs.webkit.org/show_bug.cgi?id=49513
383
384         * Scripts/webkitpy/style/checkers/cpp.py:
385         (check_for_function_lengths): Revert a comment change that I
386          accidentally made in r71986.
387         (check_pass_ptr_usage): Added the code to do the check.
388         (process_line): Added the call to check_pass_ptr_usage.
389         (CppChecker): Added the new error category.
390         * Scripts/webkitpy/style/checkers/cpp_unittest.py:
391         (CppStyleTestBase::perform_pass_ptr_check): Runs the new check for
392          testing purposes.
393         (PassPtrTest::*): The class/functions to unit test the new
394          functionality.
395
396 2010-11-14  David Levin  <levin@chromium.org>
397
398         Reviewed by Shinichiro Hamaji.
399
400         check-webkit-style function detection and the line count style checks should be separate.
401         https://bugs.webkit.org/show_bug.cgi?id=49512
402
403         * Scripts/webkitpy/style/checkers/cpp.py: Do the separation.
404         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Adjust the test to
405           call the detection function and fix line counts in two places now that
406           the code really only counts the lines in the body of the function.
407
408 2010-11-14  Andreas Kling  <kling@webkit.org>
409
410         Reviewed by Antonio Gomes.
411
412         http/tests/plugins tests print "Unhandled variable" to stderr under Qt
413         https://bugs.webkit.org/show_bug.cgi?id=33438
414
415         Stifle "Unhandled variable" warning to match what the other
416         TestNetscapePlugin does (NPP_GetValue in TestNetscapePlugIn/main.cpp
417         simply returns NPERR_GENERIC_ERROR for unhandled variables.)
418
419         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
420         (webkit_test_plugin_get_value):
421
422 2010-11-13  David Levin  <levin@chromium.org>
423
424         Reviewed by Eric Seidel.
425
426         check-webkit-style function detection crashes on functions in templates.
427         https://bugs.webkit.org/show_bug.cgi?id=49504
428
429         For example "template <bool shouldClose(const Element*)>".
430
431         * Scripts/webkitpy/style/checkers/cpp.py: If not function is detected, bail out.
432         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Modified the complex
433           function detection test to expose this issue.
434
435 2010-11-12  David Levin  <levin@chromium.org>
436
437         Reviewed by Shinichiro Hamaji.
438
439         check-webkit-style function detection doesn't detect indented functions declaractions.
440         https://bugs.webkit.org/show_bug.cgi?id=49446
441
442         Indented function declarations occur inside class definitions, so
443         they are a pretty common (and worth detecting).
444
445         * Scripts/webkitpy/style/checkers/cpp.py:
446           Changed regex to allow indentation.
447           Changed the function start detection to only happen when not in a
448           function.
449           Changed function end detection to work based on matching braces
450           instead of finding a close brace at the beginning of the line.
451           Fixed close_expression to do what it says when it doesn't find
452           the close.
453         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Indented function test.
454
455 2010-11-12  Daniel Bates  <dbates@rim.com>
456
457         Rubber-stamped by Nikolas Zimmermann.
458
459         Make do-file-rename rename files in the directory WebKit2.
460
461         This also makes the list of searched directories in do-file-rename
462         consistent with the list of searched directories in the script do-webcore-rename.
463
464         * Scripts/do-file-rename:
465
466 2010-11-12  Dirk Pranke  <dpranke@chromium.org>
467
468         Reviewed by James Robinson.
469
470         Attempt yet again to land the fix for bug 49360 (respecting
471         set-webkit-configuration). We need to handle the cases where
472         trying to run webkit-build-directory to find out where the
473         default configuration might be fails (that shows up on some
474         Chromium bots that apparently don't have perl installed).
475
476         https://bugs.webkit.org/show_bug.cgi?id=49360
477
478         * Scripts/webkitpy/layout_tests/port/config.py:
479         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Added.
480         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
481
482 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
483
484         Reviewed by Adam Barth.
485
486         webkit-patch land logs commit message twice
487         https://bugs.webkit.org/show_bug.cgi?id=49482
488         
489         Don't log the comment that's used by bugzilla.close_bug_as_fixed, since
490         that's just the commit message, which we now output following the
491         actual commit.
492
493         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
494
495 2010-11-12  Tony Chang  <tony@chromium.org>
496
497         Reviewed by Adam Barth.
498
499         run platform/chromium/plugins/script-object-invoke.html on all platforms
500         https://bugs.webkit.org/show_bug.cgi?id=49280
501
502         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
503         (testScriptObjectInvoke):
504         (pluginInvoke):
505
506 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
507
508         Reviewed by Adam Barth.
509
510         webkit-patch land-cowboy (and other commands) should display the committed SVN revision after landing a patch
511         https://bugs.webkit.org/show_bug.cgi?id=49471
512
513         Right now there's no indication of a successful commit.
514         
515         * Scripts/webkitpy/tool/commands/download_unittest.py:
516         * Scripts/webkitpy/tool/steps/commit.py:
517
518 2010-11-12  Kinuko Yasuda  <kinuko@chromium.org>
519
520         Reviewed by Ojan Vafai.
521
522         Include detailed test modifiers in results.json and enable incremental uploading for non-layout tests
523         https://bugs.webkit.org/show_bug.cgi?id=49354
524
525         Also moved/integrated the upload method from run_webkit_tests.py to json_results_generator.py.
526
527         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
528         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator.py:
529         * Scripts/webkitpy/layout_tests/layout_package/json_results_generator_unittest.py:
530         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
531
532 2010-11-12  Tony Chang  <tony@chromium.org>
533
534         Reviewed by Kent Tamura.
535
536         [chromium] Load a blank page when closing DRT WebViewHosts
537         https://bugs.webkit.org/show_bug.cgi?id=49417
538
539         In test_shell, when we close a window, we first load about:blank
540         and run GC to fire the destruction logic of the page.  In test_shell,
541         this happens in ~TestShell.  In DRT, we manage virtual windows as
542         WebViewHosts, so we need to replicate this logic in ~WebViewHost.
543
544         Managing focus is a bit trickier, since normally you would get window
545         events to reset the focus to the right window.  We do it manually
546         instead.
547
548         This fixes 2 plugin tests on DRT.
549
550         * DumpRenderTree/chromium/TestShell.cpp:
551         (TestShell::~TestShell): Move window close logic to ~WebViewHost.
552         (TestShell::closeWindow): Move widget close() to WVH destructor and
553             manage focus manually.
554         (TestShell::closeRemainingWindows): Comment typo.
555         * DumpRenderTree/chromium/WebViewHost.cpp:
556         (WebViewHost::closeWidget):
557         (invokeCloseWidget):
558         (WebViewHost::closeWidgetSoon): Close the window asynchronously (matches
559             test_shell).
560         (WebViewHost::~WebViewHost):
561         * DumpRenderTree/chromium/WebViewHost.h:
562
563 2010-11-12  David Levin  <levin@chromium.org>
564
565         Reviewed by Adam Barth.
566
567         cpp.py needs some tidying.
568         https://bugs.webkit.org/show_bug.cgi?id=49443
569
570         * Scripts/webkitpy/style/checkers/cpp.py:
571           Fix parameter ordering for iteratively_replace_matches_with_char
572           to mirror other similar functions (and improved comments, etc.)
573           Also, use a python-ism (for/else) to get rid of a variable that
574           was there to determine if something had been found in the loop.
575
576 2010-11-12  Robert Kroeger  <rjkroege@chromium.org>
577
578         Reviewed by James Robinson.
579
580         [Chromium/DRT] Make EventSendingController send complete touch events.
581         https://bugs.webkit.org/show_bug.cgi?id=49285
582
583         * DumpRenderTree/chromium/EventSender.cpp:
584         (EventSender::addTouchPoint):
585         (EventSender::updateTouchPoint):
586         (EventSender::sendCurrentTouchEvent):
587
588 2010-11-12  Mihai Parparita  <mihaip@chromium.org>
589
590         Unreviewed rollout of r71858.
591
592         Rollout out r71858 since it breaks new-run-webkit-httpd as used by the
593         NaCl tests.
594
595         * Scripts/webkitpy/layout_tests/port/config.py:
596         * Scripts/webkitpy/layout_tests/port/config_standalone.py: Removed.
597         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
598
599 2010-11-12  Benjamin Poulain  <benjamin.poulain@nokia.com>
600
601         Reviewed by Kenneth Rohde Christiansen.
602
603         [Qt] TestController::runUntil doesn't honor the timeout parameter
604         https://bugs.webkit.org/show_bug.cgi?id=48941
605
606         Implement the missing timeout of WebKitTestRunner for Qt.
607
608         A QElapsedTimer has been added to RunUntilConditionLoop in order
609         to measure how long the loop has been running. When the timer
610         is bigger or equal than the timeout, the loop ends even if the
611         condition is not met.
612
613         * WebKitTestRunner/qt/TestControllerQt.cpp:
614         (WTR::RunUntilConditionLoop::start):
615         (WTR::RunUntilConditionLoop::run):
616         (WTR::RunUntilConditionLoop::timerEvent):
617         (WTR::TestController::platformRunUntil):
618
619 2010-11-11  David Levin  <levin@chromium.org>
620
621         Reviewed by Shinichiro Hamaji.
622
623         check-webkit-style function detection doesn't handle templates with spaces.
624         https://bugs.webkit.org/show_bug.cgi?id=49427
625
626         * Scripts/webkitpy/style/checkers/cpp.py: Stripped out templates when
627           finding the function name.
628         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test have a
629           template with a space.
630
631 2010-11-11  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
632
633         Reviewed by Shinichiro Hamaji.
634
635         Enable check-webkit-style on Qt files
636         https://bugs.webkit.org/show_bug.cgi?id=48322
637
638         Adding exception for _q_... private slots and
639         ..._data() methods in tests
640
641         * Scripts/webkitpy/style/checker.py: enable Qt dirs, add exceptions
642         * Scripts/webkitpy/style/checker_unittest.py: improve unittests
643
644 2010-11-11  David Levin  <levin@chromium.org>
645
646         Reviewed by Shinichiro Hamaji.
647
648         check-webkit-style function detection doesn't work for templates and destructors.
649         https://bugs.webkit.org/show_bug.cgi?id=49425
650
651         * Scripts/webkitpy/style/checkers/cpp.py: Changed regex for templates and destructors.
652         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Made the test hit this issue.
653
654 2010-11-11  Joone Hur  <joone@kldp.org>
655
656         Reviewed by Antonio Gomes.
657
658         [GTK][DRT] Implement LayoutTestController::nodesFromRect
659         https://bugs.webkit.org/show_bug.cgi?id=46598
660
661         Support nodesFromRect in DRT
662
663         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
664         (LayoutTestController::nodesFromRect):
665
666 2010-11-11  MORITA Hajime  <morrita@google.com>
667
668         Reviewed by Kent Tamura.
669
670         editing/selection/context-menu-on-text.html fails on chromium
671         https://bugs.webkit.org/show_bug.cgi?id=45898
672
673         Enhance Chromium's EventSender::contextClick() to return
674         array of menu strings. The implementation incomplete because
675         Although actual context menu is implemented by the browser,
676         it reflects editability of underlying context thus makes
677         context-menu-on-text.html passed.
678         
679         * DumpRenderTree/chromium/EventSender.cpp:
680         (makeMenuItemStringsFor): Added.
681         (EventSender::contextClick): Now returns array of menu strings.
682         * DumpRenderTree/chromium/WebViewHost.cpp:
683         (WebViewHost::showContextMenu): Now storing given WebContextMenuData object.
684         (WebViewHost::clearContextMenuData): Added.
685         (WebViewHost::lastContextMenuData): Added.
686         * DumpRenderTree/chromium/WebViewHost.h:
687
688 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
689
690         Unreviewed, build fix.
691
692         fix build breakage caused by fix for bug 49380 (r71858).
693
694         config_standalone needs to ensure that the packages it needs are
695         in sys.path.
696
697         https://bugs.webkit.org/show_bug.cgi?id=49419
698
699         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
700
701 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
702
703         Reviewed by Adam Roben.
704
705         Fix NRWT to respect set-webkit-configuration again :(
706
707         This change fixes a typo in config.py that was causing the wrong
708         value to be read initially and us never actually looking into
709         the filesystem to get the default configuration.
710         
711         * Scripts/webkitpy/layout_tests/port/config.py:
712         * Scripts/webkitpy/layout_tests/port/config_standalone.py:
713         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
714
715 2010-11-11  Eric Seidel  <eric@webkit.org>
716
717         Reviewed by Adam Barth.
718
719         Split out Bug, Attachment and CommitterValidator from bugzilla.py
720         https://bugs.webkit.org/show_bug.cgi?id=49403
721
722         CommitterValidator really had nothing to do with bugzilla.py.
723         I've put it next to committers.py for now since its rather tied to that file.
724
725         * Scripts/webkitpy/common/config/committervalidator.py: Added.
726         * Scripts/webkitpy/common/config/committervalidator_unittest.py: Added.
727         * Scripts/webkitpy/common/net/bugzilla/__init__.py:
728         * Scripts/webkitpy/common/net/bugzilla/attachment.py: Added.
729         * Scripts/webkitpy/common/net/bugzilla/bug.py: Added.
730         * Scripts/webkitpy/common/net/bugzilla/bug_unittest.py: Added.
731         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py:
732         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py:
733         * Scripts/webkitpy/tool/bot/feeders.py:
734         * Scripts/webkitpy/tool/commands/queues.py:
735
736 2010-11-11  Dirk Pranke  <dpranke@chromium.org>
737
738         Reviewed by Eric Seidel.
739
740         fix missing 'pid' argument for http lockfile checking on win32
741         https://bugs.webkit.org/show_bug.cgi?id=49363
742
743         * Scripts/webkitpy/common/system/executive.py:
744
745 2010-11-11  Eric Seidel  <eric@webkit.org>
746
747         Rubber-stamped by Adam Barth.
748
749         Move bugzilla.py into its own module in preparation for splitting one-file-per-class
750         https://bugs.webkit.org/show_bug.cgi?id=49402
751
752         * Scripts/webkitpy/common/net/bugzilla/__init__.py: Added.
753         * Scripts/webkitpy/common/net/bugzilla/bugzilla.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla.py.
754         * Scripts/webkitpy/common/net/bugzilla/bugzilla_unittest.py: Renamed from WebKitTools/Scripts/webkitpy/common/net/bugzilla_unittest.py.
755
756 2010-11-10  Chris Guillory  <chris.guillory@google.com>
757
758         Reviewed by Chris Fleizach.
759
760         Use IAccessibleComparable to compare accessibility objects.
761         https://bugs.webkit.org/show_bug.cgi?id=49118
762
763         * DumpRenderTree/AccessibilityUIElement.cpp:
764         (AccessibilityUIElement::isEqual):
765         * DumpRenderTree/AccessibilityUIElement.h:
766         * DumpRenderTree/win/AccessibilityUIElementWin.cpp:
767         (comparableObject):
768         (AccessibilityUIElement::isEqual):
769
770 2010-11-10  Csaba Osztrogonác  <ossy@webkit.org>
771
772         Reviewed by David Hyatt.
773
774         HTML5 Ruby support should be mandatory feature
775         https://bugs.webkit.org/show_bug.cgi?id=49272
776
777         Remove Ruby as optional feature.
778
779         * Scripts/build-webkit:
780
781 2010-11-10  Adam Roben  <aroben@apple.com>
782
783         Print VC++ Express build logs in the same order that the projects are
784         built
785
786         Fixes <http://webkit.org/b/49326> print-vse-failure-logs prints logs
787         in an unhelpful order
788
789         Reviewed by Eric Seidel.
790
791         * Scripts/print-vse-failure-logs:
792         (PrintVisualStudioExpressLogs._build_order): Added. Uses
793         print-msvc-project-dependencies to figure out the order in which
794         projects are built.
795         (PrintVisualStudioExpressLogs._sort_buildlogs): Added. Sorts the logs
796         based on their build order and project name.
797         (PrintVisualStudioExpressLogs._obj_directory): Moved code to find the
798         scripts directory from here...
799         (PrintVisualStudioExpressLogs._scripts_directory): ...to here.
800         (PrintVisualStudioExpressLogs.main): Sort the logs before printing
801         them.
802
803 2010-11-10  Adam Roben  <aroben@apple.com>
804
805         Only print the interesting text from VC++ Express build logs
806
807         Fixes <http://webkit.org/b/49325> It's hard to find the interesting
808         output from print-vse-failure-logs
809
810         Reviewed by Eric Seidel.
811
812         * Scripts/print-vse-failure-logs: Removed unnecessary os.path import,
813         added newly-required imports. Removed a comment that was essentially a
814         FIXME, and is now fixed.
815         (PrintVisualStudioExpressLogs._relevant_text): Added. Uses
816         BeautifulSoup to extract the relevant text from the build log.
817         (PrintVisualStudioExpressLogs.main): Only print the relevant text.
818
819 2010-11-09  Mihai Parparita  <mihaip@chromium.org>
820
821         Reviewed by Tony Chang.
822
823         [Chromium] http/tests/security/XFrameOptions fail with Chromium DRT
824         https://bugs.webkit.org/show_bug.cgi?id=49286
825         
826         The tests in http/tests/security/XFrameOptions started to fail after
827         r71297, since it changed the WebViewHost::assignIdentifierToRequest
828         logic to always assign identifiers to requests, even if we're not going
829         to be dumping resource load callbacks at that point. These tests
830         only call LayoutTestController.dumpResourceLoadCallbacks after the
831         request is started, thus their expectations have "<unknown>" as the
832         identifier. Change WebViewHost to the previous behavior, which also
833         match the other ports'.
834
835         * DumpRenderTree/chromium/WebViewHost.cpp:
836         (WebViewHost::assignIdentifierToRequest):
837         (WebViewHost::removeIdentifierForRequest):
838
839 2010-11-08  Tony Chang  <tony@chromium.org>
840
841         Reviewed by Adam Barth.
842
843         run platform/chromium/plugins/return-npobject.html on all platforms
844         https://bugs.webkit.org/show_bug.cgi?id=49103
845
846         This tests that we can get an NPObject returned through a method on
847         an NPAPI Object.
848
849         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
850         (testCallbackReturn):
851         (pluginInvoke):
852
853 2010-11-09  Sheriff Bot  <webkit.review.bot@gmail.com>
854
855         Unreviewed, rolling out r71672.
856         http://trac.webkit.org/changeset/71672
857         https://bugs.webkit.org/show_bug.cgi?id=49276
858
859         It doesn't work on Chromium Windows bot (Requested by Ossy on
860         #webkit).
861
862         * Scripts/webkitpy/common/system/file_lock.py: Removed.
863         * Scripts/webkitpy/common/system/file_lock_unittest.py: Removed.
864         * Scripts/webkitpy/layout_tests/port/http_lock.py:
865
866 2010-11-09  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
867
868         Reviewed by Tony Chang.
869
870         [NRWT] Make http locking similar to perl implementation
871         https://bugs.webkit.org/show_bug.cgi?id=49187
872
873         * Scripts/webkitpy/common/system/file_lock.py: Added.
874         * Scripts/webkitpy/common/system/file_lock_unittest.py: Added.
875         * Scripts/webkitpy/layout_tests/port/http_lock.py:
876
877 2010-11-09  James Kozianski  <koz@chromium.org>
878
879         Reviewed by Eric Seidel.
880
881         Clean up imports for webkit-patch.
882         https://bugs.webkit.org/show_bug.cgi?id=49083
883
884         Move some imports from commands into main.py where they are used.
885         Move command imports into the __init__ of the commands module.
886
887         * Scripts/webkitpy/tool/commands/__init__.py:
888         * Scripts/webkitpy/tool/commands/download.py:
889         * Scripts/webkitpy/tool/commands/queues.py:
890         * Scripts/webkitpy/tool/main.py:
891
892 2010-11-09  Yi Shen  <yi.4.shen@nokia.com>
893
894         Reviewed by Andreas Kling.
895
896         [Qt][QtTestBrowser] Don't erase incorrect url in the Url Bar
897         https://bugs.webkit.org/show_bug.cgi?id=49047
898
899         * QtTestBrowser/mainwindow.cpp:
900         (MainWindow::setAddressUrl):
901
902 2010-11-09  Andras Becsi  <abecsi@webkit.org>
903
904         Reviewed by Csaba Osztrogonác.
905
906         Improve old-run-webkit-tests --verbose to show DumpTool opening/closing
907         https://bugs.webkit.org/show_bug.cgi?id=49252
908
909         * Scripts/old-run-webkit-tests:
910
911 2010-11-09  Csaba Osztrogonác  <ossy@webkit.org>
912
913         Reviewed by Tor Arne Vestbø.
914
915         Output of old-run-webkit-tests is buggy in verbose mode
916         https://bugs.webkit.org/show_bug.cgi?id=49249
917
918         * Scripts/old-run-webkit-tests: Redundant outputs removed.
919
920 2010-11-08  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
921
922         Reviewed by Martin Robinson.
923
924         [GTK] Link with target name set does not work
925         https://bugs.webkit.org/show_bug.cgi?id=48865
926
927         When a new page is created with a name (target=myFrame), the new
928         mainFrame could not be found because they where not stored in the
929         same PageGroup. As PageGroup are not exposed externally so the
930         simpliest solution is to use a global page group name. This also fixes
931         issue with visited link coloration across pages. After this change the
932         private function webkit_web_view_set_group_name() was no longer used
933         so it was removed completly.
934
935         * DumpRenderTree/gtk/DumpRenderTree.cpp:
936         (createWebView):
937
938 2010-11-08  John Knottenbelt  <jknotten@chromium.org>
939
940         Reviewed by Kent Tamura.
941
942         WebViewHost::reset() uses placement new.
943         https://bugs.webkit.org/show_bug.cgi?id=49069
944
945         WebViewHost is using placement destruction / new to simulate a fresh
946         WebViewHost object at the same address. This is because the WebView remains
947         open across tests and maintains a pointer to the WebViewHost.
948         This change resets member variables explictly instead of the placement new dance.
949
950         * DumpRenderTree/chromium/WebViewHost.cpp:
951         (WebViewHost::WebViewHost):
952         (WebViewHost::reset):
953
954 2010-11-08  Dirk Pranke <dpranke@chromium.org>
955
956         Reviewed by Eric Seidel.
957
958         Fix the webkit_base_dir logic in webkitpy/layout_tests/port/config.py. 
959         It turns out that NRWT can't use the code in scm.find_checkout_root()
960         because the Chromium bots don't do full checkouts of the WebKit
961         tree; they only check out subdirectories like WebKitTools/Scripts.
962         Until we can figure out a better approach for this, I've
963         restored the base_dir-detecting code from NRWT, which works in
964         any directory tree, scm or no.
965
966         This also restores the files modified in r71475 and r71474.
967
968         https://bugs.webkit.org/show_bug.cgi?id=49151
969
970         * WebKitTools/Scripts/webkitpy/layout_tests/port/config.py
971         * WebKitTools/Scripts/webkitpy/layout_tests/port/config_unittest.py
972
973 2010-11-08  Eric Seidel  <eric@webkit.org>
974
975         Reviewed by Adam Barth.
976
977         webkit-patch failure-reason explodes if a build is missing
978         https://bugs.webkit.org/show_bug.cgi?id=49195
979
980         This is likely a recent regression.  Adding a unit test for this case.
981
982         * Scripts/webkitpy/common/net/buildbot.py:
983         * Scripts/webkitpy/common/net/buildbot_unittest.py:
984
985 2010-11-08  Dan Bernstein  <mitz@apple.com>
986
987         Reviewed by Darin Adler.
988
989         Ignore the system scrollbar setting.
990
991         * DumpRenderTree/mac/DumpRenderTree.mm:
992         (resetDefaultsToConsistentValues): Set the scrollbar default.
993
994 2010-11-05  Adam Roben  <aroben@apple.com>
995
996         Make webkitpy.common.system.executive_unittest pass when running under
997         Win32 Python
998
999         Fixes <http://webkit.org/b/49033>.
1000
1001         Reviewed by Dave Levin and Eric Seidel.
1002
1003         * Scripts/webkitpy/common/system/executive.py:
1004         (Executive._run_command_with_teed_output): Pass the arguments through
1005         encode_argument_if_needed rather than using Cygwin-specific code here.
1006         (Executive.run_and_throw_if_fail): Use child_process_encoding to decode
1007         the output.
1008         (Executive.run_command): Use encode_argument_if_needed to encode the
1009         arguments and child_process_encoding to decode the output.
1010         (Executive._child_process_encoding): Returns the encoding that should be
1011         used when communicating with child processes. On Windows we use mbcs,
1012         which maps to the current code page. On all other platforms we use
1013         UTF-8.
1014         (Executive._should_encode_child_process_arguments): Returns True if
1015         arguments to child processes need to be encoded. This is currently
1016         only needed on Cygwin and Win32 Python 2.x.
1017         (Executive._encode_argument_if_needed): Encode the argument using
1018         child_process_encoding if we need to encode arguments to child
1019         processes on this platform.
1020
1021         * Scripts/webkitpy/common/system/executive_unittest.py:
1022         (never_ending_command): Added. Returns arguments to run a command that
1023         will not quit until we kill it. On Windows we use wmic, on other
1024         platforms we use yes.
1025         (ExecutiveTest.test_run_command_with_unicode): Changed to expect the
1026         mbcs encoding to be used and for output from the child processes to
1027         have been roundtripped through encode/decode on Win32 Python. When
1028         UTF-8 is the encoding the roundtripping is undetectable, but with mbcs
1029         it's possible that some characters will not be able to be converted
1030         and will be replaced by question marks; the round-tripping allows us
1031         to expect this result.
1032
1033         (ExecutiveTest.test_kill_process):
1034         (ExecutiveTest.test_kill_all):
1035         Use never_ending_command instead of invoking "yes" directly. Expect an
1036         exit code of 1 when using Win32 Python, as that's what seems to happen.
1037
1038 2010-11-08  Adam Roben  <aroben@apple.com>
1039
1040         Roll out r71532
1041
1042         It broke the build for Cygwin 1.7 installs. Cygwin 1.7's default
1043         .bashrc unsets %TEMP%, which broke copy-tools.cmd.
1044
1045         * Scripts/webkitdirs.pm:
1046
1047 2010-11-08  Tony Chang  <tony@chromium.org>
1048
1049         Reviewed by Adam Barth.
1050
1051         run platform/chromium/plugins/nested-plugin-objects.html on all platforms
1052         https://bugs.webkit.org/show_bug.cgi?id=49094
1053
1054         This tests that objects created by plugins are proplery cleaned up.
1055
1056         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1057         (testPassTestObject):
1058         (pluginInvoke):
1059         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1060         (testAllocate):
1061         (testDeallocate):
1062         (testGetProperty):
1063         (testConstruct):
1064
1065 2010-11-08  Adam Roben  <aroben@apple.com>
1066
1067         Mark Windows builds triggered from Perl as being non-interactive
1068
1069         This affects whether some of our scripts will show alerts vs. printing
1070         to the build log.
1071
1072         Fixes <http://webkit.org/b/49181> Windows build fail mysteriously when
1073         .vsprops files are updated
1074
1075         Reviewed by Steve Falkenburg.
1076
1077         * Scripts/webkitdirs.pm:
1078         (buildVisualStudioProject): Set WEBKIT_NONINTERACTIVE_BUILD to 1.
1079
1080 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1081
1082         Reviewed by Ojan Vafai.
1083
1084         Make http locking default in NRWT.
1085         https://bugs.webkit.org/show_bug.cgi?id=48053
1086
1087         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1088         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1089
1090 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1091
1092         Reviewed by Csaba Osztrogonác.
1093
1094         [NRWT] If the http lock fails we shouldn't do any locking
1095         https://bugs.webkit.org/show_bug.cgi?id=49164
1096
1097         If something goes wrong with the locking, the test should keep going.
1098
1099         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1100
1101 2010-11-08  Adam Roben  <aroben@apple.com>
1102
1103         Switch back to using kCGImageAlphaPremultipliedFirst when generating
1104         pixel dumps on Windows
1105
1106         I changed this behavior in r71418 thinking that it was required for
1107         getting plugins to show up in pixel dumps. But it doesn't seem to be
1108         necessary, and was making it impossible to compare new Windows pixel
1109         dumps with existing Windows or Mac pixel dumps (because ImageDiff won't
1110         compare an image with alpha to an image without alpha).
1111
1112         Fixes <http://webkit.org/b/49172> REGRESION (r71418): Can't compare
1113         new Windows pixel results to existing Windows or Mac results
1114
1115         Reviewed by Antti Koivisto.
1116
1117         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1118         (createBitmapContextFromWebView): Replaced kCGImageAlphaNoneSkipFirst
1119         with kCGImageAlphaPremultipliedFirst.
1120
1121 2010-11-08  Csaba Osztrogonac  <ossy@webkit.org>
1122
1123         Unreviewed, rolling out r71466.
1124         http://trac.webkit.org/changeset/71466
1125         https://bugs.webkit.org/show_bug.cgi?id=48865
1126
1127         It broke layout tests on GTK bots.
1128
1129         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1130         (createWebView):
1131
1132 2010-11-08  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1133
1134         Reviewed by Csaba Osztrogonác.
1135
1136         Enable running of Qt API tests on BuildBot
1137         https://bugs.webkit.org/show_bug.cgi?id=49004
1138
1139         * BuildSlaveSupport/build.webkit.org-config/master.cfg:
1140
1141 2010-11-08  Eric Seidel  <eric@webkit.org>
1142
1143         Reviewed by Adam Barth.
1144
1145         Add clean-review-queue command to remove closed bugs from the webkit.org/pending-review
1146         https://bugs.webkit.org/show_bug.cgi?id=49160
1147
1148         Bugzilla doesn't automatically remove r? when a bug gets closed.
1149         This script takes care of that for webkit.org.
1150
1151         * Scripts/webkitpy/common/net/bugzilla.py:
1152         * Scripts/webkitpy/tool/commands/upload.py:
1153
1154 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1155
1156         Unreviewed, rolling out r71474.
1157         http://trac.webkit.org/changeset/71474
1158         https://bugs.webkit.org/show_bug.cgi?id=48280
1159
1160         breaks chromium webkit tests
1161         https://bugs.webkit.org/show_bug.cgi?id=49151
1162
1163         * Scripts/webkitpy/layout_tests/port/base.py:
1164         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1165         * Scripts/webkitpy/layout_tests/port/config.py:
1166         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1167         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1168         * Scripts/webkitpy/layout_tests/port/webkit.py:
1169         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1170         * Scripts/webkitpy/tool/mocktool.py:
1171
1172 2010-11-07  Fumitoshi Ukai  <ukai@chromium.org>
1173
1174         Unreviewed, rolling out r71475.
1175         http://trac.webkit.org/changeset/71475
1176  
1177         breaks chromium webkit tests
1178         https://bugs.webkit.org/show_bug.cgi?id=49151
1179
1180         * Scripts/webkitpy/common/newstringio.py: Removed.
1181         * Scripts/webkitpy/common/newstringio_unittest.py: Removed.
1182         * Scripts/webkitpy/common/system/executive_mock.py: Removed.
1183         * Scripts/webkitpy/common/system/filesystem_mock.py: Removed.
1184         * Scripts/webkitpy/layout_tests/port/config_mock.py: Removed.
1185
1186 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1187
1188         Unreviewed, build fix.
1189
1190         Add files inexplicably not committed in r71474 as part of the
1191         fix for bug 48280.
1192
1193         * Scripts/webkitpy/common/newstringio.py: Added.
1194         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1195         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1196         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1197         * Scripts/webkitpy/layout_tests/port/config_mock.py: Added.
1198
1199 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1200
1201         Reviewed by Eric Siedel.
1202
1203         new-run-webkit-tests: update port/base and port/webkit to use the
1204         new FileSystem and Config abstractions, pulling more logic out of
1205         the base Port classes into separate, mockable objects.
1206
1207         Also create a MockFileSystem object, a MockConfig object, move
1208         MockExecutive into common/system to be next to executive, and
1209         update the config object to use a FileSystem.
1210
1211         https://bugs.webkit.org/show_bug.cgi?id=48280
1212
1213         * Scripts/webkitpy/common/newstringio.py: Added.
1214         * Scripts/webkitpy/common/newstringio_unittest.py: Added.
1215         * Scripts/webkitpy/common/system/executive_mock.py: Added.
1216         * Scripts/webkitpy/common/system/filesystem_mock.py: Added.
1217         * Scripts/webkitpy/layout_tests/port/base.py:
1218         * Scripts/webkitpy/layout_tests/port/base_unittest.py:
1219         * Scripts/webkitpy/layout_tests/port/config.py:
1220         * Scripts/webkitpy/layout_tests/port/config_mock.py:
1221         * Scripts/webkitpy/layout_tests/port/config_unittest.py:
1222         * Scripts/webkitpy/layout_tests/port/google_chrome_unittest.py:
1223         * Scripts/webkitpy/layout_tests/port/webkit.py:
1224         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1225         * Scripts/webkitpy/tool/mocktool.py:
1226
1227 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1228
1229         Unreviewed, build breakage.
1230
1231         Apparently I uploaded the wrong version of the file to fix 49122
1232         and neither Eric or I noticed - it was missing a dirname() call.
1233         Fixing ...
1234
1235         https://bugs.webkit.org/show_bug.cgi?id=49122
1236
1237         * Scripts/webkitpy/common/checkout/scm.py:
1238
1239 2010-11-06  Dirk Pranke  <dpranke@chromium.org>
1240
1241         Reviewed by Eric Seidel.
1242
1243         webkitpy/tool/* unittests change cwd and don't clean up properly
1244
1245         https://bugs.webkit.org/show_bug.cgi?id=49122
1246
1247         * Scripts/webkitpy/common/checkout/scm.py:
1248
1249 2010-11-05  Chris Marrin  <cmarrin@apple.com>
1250
1251         Reviewed by Simon Fraser.
1252
1253         suspendAnimations/resumeAnimations not present in WebKit2
1254         https://bugs.webkit.org/show_bug.cgi?id=49109
1255
1256         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
1257         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
1258         (WTR::LayoutTestController::suspendAnimations):
1259         (WTR::LayoutTestController::resumeAnimations):
1260         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
1261
1262 2010-11-05  Tony Chang  <tony@chromium.org>
1263
1264         Reviewed by David Levin.
1265
1266         cleanup style in TestNetscapePlugIn/PluginObject.cpp
1267         https://bugs.webkit.org/show_bug.cgi?id=49044
1268
1269         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1270         (getPluginClass):
1271         (pluginGetProperty):
1272         (pluginSetProperty):
1273         (variantToIdentifier):
1274         (testCallback):
1275         (getURL):
1276         (getURLNotify):
1277         (testInvokeDefault):
1278         (testGetProperty):
1279         (testDocumentOpen):
1280         (testWindowOpen):
1281         (pluginInvoke):
1282         (handleCallback):
1283         (notifyStream):
1284         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
1285
1286 2010-11-05  Eric Seidel  <eric@webkit.org>
1287
1288         Reviewed by David Levin.
1289
1290         Add QueueStatusServer/__init__.py so others can run the QueueStatusServer tests
1291         https://bugs.webkit.org/show_bug.cgi?id=49032
1292
1293         I wrote this file as part of bug 47847, but I forgot to commit it.
1294         No one else noticed it missing because test-webkitpy knows how
1295         to recover in the case where it can't import QueueStatusServer
1296         (which generally occurs due to not having installed the AppEngine SDK).
1297
1298         * QueueStatusServer/__init__.py: Added.
1299         * QueueStatusServer/model/workitems_unittest.py:
1300          - Remove a test which fails.  This was probably landed (by me)
1301            from my other machine, which since this __init__.py was missing
1302            I never noticed the failure and landed this invalid test.
1303            Sadly we can't really test remove_work_item as it depends
1304            on .key() working.  .key() will throw unless the object
1305            has already been saved it seems.
1306            This may be a mis-design in our remove_work_item implementation,
1307            but for now, just removing the test.
1308
1309 2010-11-04  Adam Roben  <aroben@apple.com>
1310
1311         Teach check-webkit-style about TestNetscapePlugIn
1312
1313         Fixes <http://webkit.org/b/49030> check-webkit-style is ignorant of
1314         TestNetscapePlugIn's coding conventions
1315
1316         Reviewed by John Sullivan.
1317
1318         * Scripts/webkitpy/style/checker.py: Lump TestNetscapePlugIn in with
1319         WebKitAPITest and TestWebKitAPI in having different include order and
1320         naming conventions than WebCore. Added some comments about why the
1321         exceptions exist.
1322
1323 2010-11-04  Adam Roben  <aroben@apple.com>
1324
1325         Add a test that shows that windowed plugins are able to paint
1326
1327         Somehow we never had a test for this before.
1328
1329         Fixes <http://webkit.org/b/49029> Should add a test that shows
1330         windowed plugins can paint
1331
1332         Reviewed by Jon Honeycutt.
1333
1334         * DumpRenderTree/TestNetscapePlugIn/Tests/win/DrawsGradient.cpp: Added.
1335         (DrawsGradient::DrawsGradient):
1336         (DrawsGradient::wndProc): We handle the WM_PAINT and WM_PRINTCLIENT messages.
1337
1338         (DrawsGradient::onPaint):
1339         (DrawsGradient::onPrintClient):
1340         These both just call through to paint.
1341
1342         (DrawsGradient::paint): Fills our client area with some gradients.
1343
1344         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj: Link
1345         against Msimg32.lib for ::GradientFill and added DrawsGradient.
1346
1347         * DumpRenderTree/win/PixelDumpSupportWin.cpp:
1348         (createBitmapContextFromWebView): Use WM_PRINT instead of
1349         WM_PRINTCLIENT so that ::DefWindowProc will send
1350         WM_PRINT/WM_PRINTCLIENT messages to the WebView's child windows.
1351         Replaced kCGImageAlphaPremultipledFirst with kCGImageAlphaNoneSkipFirst
1352         because GDI doesn't support alpha and kCGImageBitmapOrder32Little with
1353         kCGImageBitmapOrder32Host because, who knows, maybe someday Windows
1354         will run on a big-endian platform.
1355
1356 2010-11-04  Adam Roben  <aroben@apple.com>
1357
1358         Extract much of NPNInvalidateRectInvalidatesWindow's code into a
1359         WindowedPluginTest base class
1360
1361         The base class takes care of subclassing the plugin's window so that a
1362         custom WNDPROC is called. This will make it easier to write tests that
1363         need to handle window messages.
1364
1365         Fixes <http://webkit.org/b/49028> It's hard to write a PluginTest with
1366         a custom WNDPROC
1367
1368         Reviewed by Jon Honeycutt.
1369
1370         * DumpRenderTree/TestNetscapePlugIn/Tests/win/NPNInvalidateRectInvalidatesWindow.cpp:
1371         (NPNInvalidateRectInvalidatesWindow::NPNInvalidateRectInvalidatesWindow):
1372         (NPNInvalidateRectInvalidatesWindow::NPP_SetWindow):
1373         (NPNInvalidateRectInvalidatesWindow::wndProc):
1374         (NPNInvalidateRectInvalidatesWindow::testInvalidateRect):
1375         Moved code from here to WindowedPluginTest. Changed to use window()
1376         instead of m_window.
1377
1378         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1379         Added WindowedPluginTest and added TestNetscapePlugIn/win to the
1380         include path. Also reordered the include path to make a little more
1381         sense and simplified the entry that added TestNetscapePlugIn itself to
1382         the include path.
1383
1384         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.cpp: Added.
1385         (WindowedPluginTest::WindowedPluginTest):
1386         (WindowedPluginTest::NPP_SetWindow):
1387         (WindowedPluginTest::staticWndProc):
1388         * DumpRenderTree/TestNetscapePlugIn/win/WindowedPluginTest.h: Added.
1389         (WindowedPluginTest::window):
1390         Code came from NPNInvalidateRectInvalidatesWindow.
1391
1392 2010-11-04  Adam Roben  <aroben@apple.com>
1393
1394         Add a plugin test to show that windowed plugins are clipped correctly
1395
1396         Fixes <http://webkit.org/b/49024> <rdar://problem/8487847> Windowed
1397         plugins aren't clipped in WebKit2 on Windows
1398
1399         Reviewed by Jon Honeycutt.
1400
1401         * DumpRenderTree/TestNetscapePlugIn/Tests/win/WindowRegionIsSetToClipRect.cpp: Added.
1402         (WindowRegionIsSetToClipRect::WindowRegionIsSetToClipRect): Initialize members.
1403         (WindowRegionIsSetToClipRect::NPP_SetWindow): Check that our window
1404         region matches the clip rect we know we should have based on
1405         window-region-is-set-to-clip-rect.html, and check that our window class
1406         doesn't have the CS_PARENTDC style.
1407
1408         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1409         Added WindowRegionIsSetToClipRect.
1410
1411 2010-11-05  Alejandro G. Castro  <alex@igalia.com>
1412
1413         Reviewed by Martin Robinson.
1414
1415         [GTK] Avoid font hinting in the DRT
1416         https://bugs.webkit.org/show_bug.cgi?id=48548
1417
1418         Change the settings to avoid font hinting, it was causing
1419         different results depending on the distribution.
1420
1421         * DumpRenderTree/gtk/DumpRenderTree.cpp:
1422         (initializeGtkFontSettings):
1423         * DumpRenderTree/gtk/fonts/fonts.conf:
1424
1425 2010-11-05  Eric Seidel  <eric@webkit.org>
1426
1427         Reviewed by Adam Barth.
1428
1429         build-webkit spams system.log by repeatedly checking for PBXProductDirectory in com.apple.Xcode
1430         https://bugs.webkit.org/show_bug.cgi?id=49051
1431
1432         This is a speculative fix.  The unit tests cover these methods, however
1433         I don't know if this will fully stop the system.log spam.
1434
1435         * Scripts/webkitpy/layout_tests/port/base.py:
1436         * Scripts/webkitpy/layout_tests/port/webkit.py:
1437
1438 2010-11-05  Eric Seidel  <eric@webkit.org>
1439
1440         Reviewed by Adam Barth.
1441
1442         Add basic support for showing bot id on /queue-status/ pages
1443         https://bugs.webkit.org/show_bug.cgi?id=49037
1444
1445         This support is really simple.  Eventually we'll want to
1446         show the bot id in the lock table too, but we don't have
1447         that information stored in the server yet.
1448
1449         * QueueStatusServer/handlers/queuestatus.py:
1450         * QueueStatusServer/handlers/queuestatus_unittest.py: Added.
1451         * QueueStatusServer/templates/includes/singlequeuestatus.html:
1452
1453 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1454
1455         Reviewed by Adam Barth.
1456
1457         Create a filesystem wrapper that we can use to enforce
1458         particular conventions and use for mocking and dependency
1459         injection down the line.
1460
1461         https://bugs.webkit.org/show_bug.cgi?id=48144
1462
1463         * Scripts/webkitpy/common/system/filesystem.py: Added.
1464         * Scripts/webkitpy/common/system/filesystem_unittest.py: Added.
1465
1466 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1467
1468         Reviewed by Tony Chang.
1469
1470         run_webkit_tests_unittest fails under Python 2.5
1471         https://bugs.webkit.org/show_bug.cgi?id=49043
1472         
1473         Switch from itertools.chain.from_iterable (which was added in 2.6)
1474         to using itertools.chain directly.
1475
1476         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1477
1478 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1479
1480         Unreviewed fix to webkit-patch rebaseline-server so that it runs under
1481         Python 2.5 (it needs an import to use the with statement).
1482
1483         * Scripts/webkitpy/tool/commands/rebaselineserver.py:
1484
1485 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1486
1487         Reviewed by Adam Barth.
1488
1489         new-run-webkit-tests: split out webkit-specific configuration stuff into a new module
1490
1491         The current NRWT code has webkit-specific configuration code (like
1492         _script_path, default configuration, etc.) mixed in with
1493         layout-test-specific stuff in port/base. The configuration code
1494         should be split out into a separate module for easier mocking,
1495         testing, and isolation.
1496
1497         https://bugs.webkit.org/show_bug.cgi?id=48264
1498
1499         * Scripts/webkitpy/layout_tests/port/config.py: Added.
1500         * Scripts/webkitpy/layout_tests/port/config_unittest.py: Added.
1501
1502 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1503
1504         Reviewed by Tony Chang.
1505
1506         Rebaseline server: initial framework
1507         https://bugs.webkit.org/show_bug.cgi?id=48892
1508         
1509         Adds the basic framework for the rebaseline server (details at 
1510         http://webkit.org/b/47761). Includes the rebaseline-server webkit-patch
1511         command, which starts an HTTP server that can serve static files or
1512         invoke handler methods on a class.
1513
1514         * Scripts/webkitpy/tool/commands/__init__.py:
1515         * Scripts/webkitpy/tool/commands/data/rebaselineserver/index.html: Added.
1516         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.css: Added.
1517         * Scripts/webkitpy/tool/commands/data/rebaselineserver/main.js: Added.
1518         * Scripts/webkitpy/tool/commands/rebaselineserver.py: Added.
1519
1520 2010-11-04  Dirk Pranke  <dpranke@chromium.org>
1521
1522         Reviewed by James Robinson.
1523
1524         new-run-webkit-tests wasn't using DRT by default for
1525         --platform chromium-gpu
1526
1527         The default value was set to False instead of None, which meant
1528         that the platform specific logic wasn't firing to change the
1529         value to True (b/c we were afraid we'd be overriding the user
1530         preference).
1531
1532         https://bugs.webkit.org/show_bug.cgi?id=49038
1533
1534         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1535
1536 2010-11-04  Tony Chang  <tony@chromium.org>
1537
1538         Reviewed by Adam Barth.
1539
1540         make platform/chromium/plugins/multiple-plugins.html pass on all platforms
1541         https://bugs.webkit.org/show_bug.cgi?id=49026
1542
1543         Have the testObject.bar property return the string "bar".  This lets
1544         us run and pass platform/chromium/plugins/multiple-plugins.html on all
1545         platforms.
1546
1547         * DumpRenderTree/TestNetscapePlugIn/TestObject.cpp:
1548         (testGetProperty):
1549
1550 2010-11-04  Mihai Parparita  <mihaip@chromium.org>
1551
1552         Reviewed by Tony Chang.
1553
1554         NRWT doesn't have good test coverage for --run-chunk, --batch-size, --run-part, etc.
1555         https://bugs.webkit.org/show_bug.cgi?id=48878
1556         
1557         Add get_tests_run so that it's easy to see which tests get run (and with
1558         what batching) for a given flag combination. Flesh out the various
1559         test cases that have FIXMEs.
1560         
1561         Also fixes an off-by-one error (batch sizes were one larger than 
1562         expected) and makes --run-part also have wraparound behavior, like
1563         --run-chunk.
1564
1565         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1566         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1567         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py:
1568
1569 2010-11-04  Erik Arvidsson  <arv@chromium.org>
1570
1571         Reviewed by Dimitri Glazkov.
1572
1573         Support box-sizing without the vendor prefix
1574         https://bugs.webkit.org/show_bug.cgi?id=36713
1575
1576         Based on patch by Peter Beverloo <peter@lvp-media.com>
1577
1578         * iExploder/htdocs/cssproperties.in: Change -webkit-box-sizing to box-sizing.
1579
1580 2010-11-04  Csaba Osztrogonác  <ossy@webkit.org>
1581
1582         Unreviewed rollout r71340, because it broke Chromium Windows bot.
1583
1584         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1585         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1586
1587 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1588
1589         Reviewed by Ojan Vafai.
1590
1591         Make http locking default in NRWT.
1592         https://bugs.webkit.org/show_bug.cgi?id=48053
1593
1594         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
1595         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
1596
1597 2010-11-04  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
1598
1599         Reviewed by Eric Seidel.
1600
1601         [NRWT] Clear invalid http locks on Windows platform as well
1602         https://bugs.webkit.org/show_bug.cgi?id=48515
1603
1604         * Scripts/webkitpy/common/system/executive.py:
1605         * Scripts/webkitpy/common/system/executive_unittest.py:
1606         * Scripts/webkitpy/layout_tests/port/http_lock.py:
1607
1608 2010-11-02  Adam Roben  <aroben@apple.com>
1609
1610         Reduce our dependence on coreutils when running Python tests
1611
1612         This patch introduces versions of the standard echo and cat utilities
1613         implemented in Python. They are probably missing features of their
1614         coreutils equivalents, but they can do what's necessary for our Python
1615         tests. This is useful on Windows, where these utilities typically
1616         aren't available.
1617
1618         Fixes <http://webkit.org/b/48883> executive_unittest relies on echo
1619         and cat utilities from coreutils, which are not present on Windows
1620
1621         Reviewed by Eric Seidel.
1622
1623         * Scripts/webkitpy/common/system/executive_unittest.py: Changed to use
1624         our Python-based echo and cat.
1625
1626         * Scripts/webkitpy/common/system/fileutils.py: Added.
1627         (make_stdout_binary): On Windows, puts sys.stdout into binary mode so
1628         that \n won't be translated into \r\n. I couldn't think of a good way
1629         to test this directly without touching the filesystem, but it is tested
1630         indirectly by echo_unittest.
1631
1632         * Scripts/webkitpy/test/cat.py: Added.
1633         (command_arguments): Returns a list for invoking cat with the given arguments.
1634         (main): Acts like a simplified version of the coreutils cat utility.
1635
1636         * Scripts/webkitpy/test/cat_unittest.py: Added.
1637         (CatTest.assert_cat): Runs cat with the given input and ensures the
1638         output matches the input.
1639         (CatTest.test_basic): Performs a simple test of cat.
1640         (CatTest.test_no_newline): Tests what happens when the input string
1641         doesn't have a trailing newline.
1642         (CatTest.test_unicode): Tests passing a unicode string to cat.
1643         (CatTest.test_as_command): Tests running cat as a separate command.
1644
1645         * Scripts/webkitpy/test/echo.py: Added.
1646         (command_arguments): Returns a list for invoking echo with the given arguments.
1647         (main): Acts like a simplified version of the coreutils echo utility.
1648
1649         * Scripts/webkitpy/test/echo_unittest.py: Added.
1650         (EchoTest.test_basic): Performs a simple test of echo.
1651         (EchoTest.test_no_newline): Tests passing -n to echo to suppress the
1652         trailing newline.
1653         (EchoTest.test_unicode): Tests passing unicode and non-unicode strings
1654         to echo.
1655         (EchoTest.test_argument_order): Tests what happens when -n is not the
1656         first argument.
1657         (EchoTest.test_empty_arguments): Tests what happens when you pass [] to
1658         echo.main.
1659         (EchoTest.test_no_arguments): Tests what happens when you call
1660         echo.main with no arguments.
1661         (EchoTest.test_as_command): Tests running echo as a separate command.
1662
1663 2010-11-04  Renata Hodovan  <reni@inf.u-szeged.hu>
1664
1665         Unreviewed: Add myself to the list of Committers.
1666
1667         * Scripts/webkitpy/common/config/committers.py:
1668
1669 2010-11-04  Andreas Kling  <kling@webkit.org>
1670
1671         Reviewed by Kenneth Rohde Christiansen.
1672
1673         [Qt] Hook into QWebPage::scrollRequested for resizes-to-contents mode
1674
1675         In RTC mode the QGraphicsWebView item is the size of the contents,
1676         scrolling works a bit differently  (we need to react to scrollRequested.)
1677
1678         Normally QGraphicsView will replay the last mouse event when scrolling,
1679         so to prevent WebKit from getting confused by this we temporarily make
1680         the QGraphicsView non-interactive.
1681
1682         * QtTestBrowser/webview.cpp:
1683         (WebViewGraphicsBased::setPage):
1684         (WebViewGraphicsBased::scrollRequested):
1685         * QtTestBrowser/webview.h:
1686
1687 2010-11-04  Tor Arne Vestbø  <tor.arne.vestbo@nokia.com>
1688
1689         Reviewed by Andreas Kling.
1690
1691         Use OS(MAC_OS_X) rather than PLATFORM(MAC) in TestControllerQt
1692
1693         PLATFORM(MAC) is not defined for the Qt port, as it refers to
1694         the Mac-port, not the Mac OS X operating system.
1695
1696         * WebKitTestRunner/qt/TestControllerQt.cpp:
1697         (WTR::isExistingLibrary):
1698
1699 2010-11-04  Leonid Ebril  <leonid.ebril@nokia.com>
1700
1701         Reviewed by Andreas Kling.
1702
1703         [Qt] Adding iphone user agent string the user agent list for QtTestBrowser
1704         https://bugs.webkit.org/show_bug.cgi?id=48869
1705
1706         * QtTestBrowser/useragentlist.txt:
1707
1708 2010-11-03  Adam Roben  <aroben@apple.com>
1709
1710         Always use uppercase drive names in strings returned by abspath_to_uri
1711
1712         Some versions of cygpath use lowercase drive letters while others use
1713         uppercase, which makes it hard to test the output of code that uses
1714         cygpath.
1715
1716         Fixes <http://webkit.org/b/48914> webkitpy.common.system.path_unittest
1717         fails with Cygwin 1.5
1718
1719         Reviewed by Eric Seidel.
1720
1721         * Scripts/webkitpy/common/system/path.py:
1722         (cygpath): Updated the docstring to indicate that only absolute paths
1723         should be passed for now (though relative paths will work fine).
1724         (_Cygpath.convert): Upper-case the first letter of the converted Windows path.
1725
1726 2010-11-03  George Guo  <George.Guo@Nokia.com>
1727
1728         Reviewed by Andreas Kling.
1729
1730         [Qt] QtTestBrowser : set mmp rule pageddata in Symbian
1731         https://bugs.webkit.org/show_bug.cgi?id=48767
1732     
1733         Paging is needd on Symbian devices to support benchmarks tests like
1734         dromaeo.com and Celtic Kane that need a lot of memory to run  
1735
1736         * QtTestBrowser/QtTestBrowser.pro:
1737
1738 2010-11-03  Jenn Braithwaite  <jennb@chromium.org>
1739
1740         Reviewed by Dmitry Titov.
1741
1742         Chromium: Update resource tracking when moving a frame between documents
1743         https://bugs.webkit.org/show_bug.cgi?id=48363
1744
1745         * DumpRenderTree/chromium/WebViewHost.cpp:
1746         (WebViewHost::assignIdentifierToRequest):
1747         Always put resource id in map so we can make assumptions about its
1748         presence.
1749         (WebViewHost::removeIdentifierForRequest):
1750         Added.
1751         (WebViewHost::didFinishResourceLoad):
1752         (WebViewHost::didFailResourceLoad):
1753         Check existence of resource id before removing from map.
1754         * DumpRenderTree/chromium/WebViewHost.h:
1755
1756 2010-11-03  Victor Wang  <victorw@chromium.org>
1757
1758         Reviewed by Adam Barth.
1759
1760         [Chromium] update buildbot names in chromium rebaseline tool.
1761
1762         https://bugs.webkit.org/show_bug.cgi?id=48881
1763
1764         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1765
1766 2010-11-03  Darin Adler  <darin@apple.com>
1767
1768         Updated Xcode projects by opening them with Xcode 3.2.4.
1769         Updated svn:ignore for Xcode projects.
1770
1771         * MiniBrowser/MiniBrowser.xcodeproj: Added property svn:ignore.
1772         * TestWebKitAPI/TestWebKitAPI.xcodeproj: Added property svn:ignore.
1773         * WebKitLauncher/WebKitLauncher.xcodeproj: Modified property svn:ignore.
1774         * WebKitLauncher/WebKitLauncher.xcodeproj/project.pbxproj: Updated with Xcode 3.2.4.
1775
1776 2010-11-03  Adam Roben  <aroben@apple.com>
1777
1778         Stop waiting for messages from the web process after a timeout elapses
1779
1780         TestController::runUntil now takes a timeout parameter that specifies
1781         how long to wait for the condition to become true. WebKitTestRunner
1782         takes a --timeout flag that specifies how long the timeout should be.
1783         run-webkit-tests passes this flag to WebKitTestRunner so its timeout
1784         will be similar to run-webkit-tests.
1785
1786         Fixes <http://webkit.org/b/43047> <rdar://problem/8365833>
1787         WebKitTestRunner waits forever if the web process crashes
1788
1789         Reviewed by Darin Adler and Anders Carlsson.
1790
1791         * Scripts/old-run-webkit-tests:
1792         (top level): Moved the GuardMalloc timeout adjustment here from
1793         readFromDumpToolWithTimer.
1794         (openDumpTool): Make WTR use a timeout similar to but slightly shorter
1795         than the one that was specified on the command line.
1796
1797         * WebKitTestRunner/TestController.cpp:
1798         (WTR::TestController::TestController): Initialize our timeout values.
1799         (WTR::TestController::initialize): Parse the --timeout flag and use it
1800         to modify our timeout values.
1801         (WTR::TestController::resetStateToConsistentValues): Changed to use a
1802         short timeout while waiting for the web process to reset and to return
1803         a boolean indicating whether we were able to reset the web process.
1804         Uses a 5-second timeout while waiting for the process to be reset.
1805         (WTR::TestController::runTest): Changed to return a boolean indicating
1806         whether we were able to reset the web process (and thus run the test).
1807         (WTR::TestController::runUntil): Call through to platformRunUntil.
1808
1809         (WTR::TestController::runTestingServerLoop):
1810         (WTR::TestController::run):
1811         Changed to bail if any test can't be run. This will cause the process
1812         to exit. (Unfortunately this will make run-webkit-tests think we
1813         crashed; see <http://webkit.org/b/48943>.)
1814
1815         * WebKitTestRunner/TestController.h: Added platformRunUntil,
1816         m_longTimeout, and m_shortTimeout.
1817
1818         * WebKitTestRunner/TestInvocation.cpp:
1819         (WTR::TestInvocation::invoke): Use a short timeout when waiting for the
1820         initial response and a long timeout when waiting for the test to
1821         complete. Check whether runUntil timed out and print an error message
1822         if so.
1823
1824         * WebKitTestRunner/mac/TestControllerMac.mm:
1825         (WTR::TestController::platformRunUntil): Renamed from runUntil. Pass
1826         [NSDate distantPast] to -[NSRunLoop runMode:beforeDate:] so that we
1827         won't block waiting for the run loop. Only loop until the timeout
1828         elapses.
1829
1830         * WebKitTestRunner/qt/TestControllerQt.cpp:
1831         (WTR::TestController::platformRunUntil): Renamed from runUntil. Added a
1832         FIXME about honoring the timeout.
1833
1834         * WebKitTestRunner/win/TestControllerWin.cpp:
1835         (WTR::TestController::platformRunUntil): Renamed from runUntil. Use
1836         ::MsgWaitForMultipleObjectsEx to implement the timeout. Changed to use
1837         ::PeekMessageW so that we don't block waiting for messages to become
1838         available.
1839
1840 2010-11-03  Adam Roben  <aroben@apple.com>
1841
1842         Add a plugin test that evaluates JS after removing the plugin element
1843         from the document
1844
1845         This test replaces platform/win/plugins/plugin-delayed-destroy.html.
1846         That test was made to prevent a crash very similar to this one, but
1847         unfortunately tested only the mechanism that prevented the crash and
1848         not whether the crash itself was prevented. Since WebKit2 uses a
1849         different mechanism to prevent the crash, the test was failing even
1850         though WebKit2 was not vulnerable to the crash. This new test crashes
1851         if there is no mechanism in place to prevent it and passes in both
1852         WebKit1 and WebKit2.
1853
1854         Fixes <http://webkit.org/b/46711> <rdar://problem/8485903>
1855         platform/win/plugins/plugin-delayed-destroy.html fails in WebKit2
1856
1857         Reviewed by Anders Carlsson.
1858
1859         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
1860         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
1861         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
1862         * GNUmakefile.am:
1863         Added new file.
1864
1865         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
1866         (pluginDeallocate): Make sure we delete the PluginTest object. This
1867         prevents a leak and also allows us to test the crash.
1868
1869         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
1870         (PluginTest::executeScript): Made this into a non-static member
1871         function.
1872
1873         (PluginTest::waitUntilDone):
1874         (PluginTest::notifyDone):
1875         Updated for changes to executeScript.
1876
1877         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h: Added executeScript.
1878
1879         * DumpRenderTree/TestNetscapePlugIn/Tests/EvaluateJSAfterRemovingPluginElement.cpp: Added.
1880         (EvaluateJSAfterRemovingPluginElement::EvaluateJSAfterRemovingPluginElement):
1881         Initialize ourselves and tell the test harness to wait.
1882         (EvaluateJSAfterRemovingPluginElement::NPP_DestroyStream): Remove our
1883         plugin element from the document, then execute some JavaScript. If
1884         WebKit does not have appropriate mechanisms in place, we'll be
1885         destroyed inside the first call to executeScript and crash on the
1886         second call.
1887
1888 2010-11-02  Stephen White  <senorblanco@chromium.org>
1889
1890         Reviewed by Tony Chang.
1891
1892         [chromium] Fix LayoutTestController UMRs.
1893         https://bugs.webkit.org/show_bug.cgi?id=48872
1894
1895         * DumpRenderTree/chromium/LayoutTestController.cpp:
1896         (LayoutTestController::LayoutTestController):
1897
1898 2010-11-03  Kent Tamura  <tkent@chromium.org>
1899
1900         Reviewed by Dimitri Glazkov.
1901
1902         REGRESSION: rebaseline-chromium-webkit-tests uses non-zero tolerance for
1903         image dup detection
1904         https://bugs.webkit.org/show_bug.cgi?id=48744
1905
1906         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests.py:
1907          - Make a function for option parsing for ease of test
1908          - Set 0 to options.tolerance
1909         * Scripts/webkitpy/layout_tests/rebaseline_chromium_webkit_tests_unittest.py:
1910          - Add a test for this change
1911
1912 2010-11-02  Gyuyoung Kim  <gyuyoung.kim@samsung.com>
1913
1914         Unreviewed: Add myself to the list of Committers.
1915
1916         * Scripts/webkitpy/common/config/committers.py:
1917
1918 2010-11-02  Anders Carlsson  <andersca@apple.com>
1919
1920         Fix build.
1921
1922         * TestWebKitAPI/PlatformUtilities.cpp:
1923         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
1924
1925 2010-11-02  Sam Weinig  <sam@webkit.org>
1926
1927         Reviewed by Anders Carlsson.
1928
1929         Add a way to send startup messages on the context which can be posted when a process launches
1930         <rdar://problem/8617928>
1931         https://bugs.webkit.org/show_bug.cgi?id=48838
1932
1933         * MiniBrowser/mac/WebBundle/WebBundleMain.m:
1934         (WKBundleInitialize):
1935         * TestWebKitAPI/InjectedBundleController.cpp:
1936         (TestWebKitAPI::InjectedBundleController::initialize):
1937         (TestWebKitAPI::InjectedBundleController::didReceiveMessage):
1938         * TestWebKitAPI/InjectedBundleController.h:
1939         * TestWebKitAPI/InjectedBundleMain.cpp:
1940         (WKBundleInitialize):
1941         * TestWebKitAPI/PlatformUtilities.cpp:
1942         (TestWebKitAPI::Util::createContextForInjectedBundleTest):
1943         * WebKitTestRunner/InjectedBundle/InjectedBundleMain.cpp:
1944         (WKBundleInitialize):
1945         Update implementations of WKBundleInitialize to take an initial userData
1946         argument.  Change TestWebKitAPI to use the new initial userData to initialize
1947         each test's bundle.
1948
1949 2010-11-02  Benjamin Kalman  <kalman@google.com>
1950
1951         Reviewed by Ojan Vafai.
1952
1953         new-run-webkit-tests doesn't strip "LayoutTests/" from prefix, unlike old-run-webkit-tests
1954         https://bugs.webkit.org/show_bug.cgi?id=48794
1955
1956         * Scripts/webkitpy/layout_tests/run_webkit_tests.py: Strip the "LayoutTests/" prefix from test argument paths.
1957
1958 2010-11-02  Adam Roben  <aroben@apple.com>
1959
1960         Skip webkitpy.layout_tests.run_webkit_tests_unittest.MainTest on Cygwin
1961         Python 2.5.x
1962
1963         It is known to hang on that version of Python. See
1964         <http://webkit.org/b/48614>.
1965
1966         Reviewed by Adam Barth.
1967
1968         * Scripts/webkitpy/layout_tests/run_webkit_tests_unittest.py: Use
1969         skip_if to skip MainTest on Cygwin Python 2.5.x.
1970
1971         * Scripts/webkitpy/test/skip.py: Added.
1972         (skip_if): If the passed-in condition is false, find all the test_*
1973         methods of the given class and replace them with a function that just
1974         logs that we're skipping these tests. This is loosely based on the
1975         unittest.skip_if decorator added in Python 3.1.
1976         (_skipped_method): Returns a function that just logs that the tests are
1977         being skipped. This is used in place of the actual test_* functions.
1978
1979         * Scripts/webkitpy/test/skip_unittest.py: Added.
1980         (SkipTest.setUp): Create a logger for ourselves and give it a handler
1981         that logs to a private stream.
1982         (SkipTest.tearDown): Reset the logger.
1983         (SkipTest.create_fixture_class): Returns a class that we use to test
1984         skip_if. It has a single test_* method, test_foo, that just calls
1985         through to a callback.
1986         (SkipTest.foo_callback): Record that test_foo was called.
1987         (SkipTest.test_skip_if_false): Pass skip_if a False condition and test
1988         that test_foo does get called.
1989         (SkipTest.test_skip_if_true): Pass skip_if a True condition and test
1990         that test_foo does not get called and the appropriate message gets
1991         logged.
1992
1993 2010-11-02  Adam Barth  <abarth@webkit.org>
1994
1995         Reviewed by Eric Seidel.
1996
1997         webkit-patch should tell check-webkit-style which files were changed so
1998         check-webkit-style doesn't have to stat the whole working copy again
1999         https://bugs.webkit.org/show_bug.cgi?id=48792
2000
2001         * Scripts/webkitpy/tool/mocktool.py:
2002         * Scripts/webkitpy/tool/steps/checkstyle.py:
2003
2004 2010-11-02  Robert Kroeger  <rjkroege@chromium.org>
2005
2006         Reviewed by James Robinson.
2007
2008         [Chromium/DRT] Make EventSendingController honour leapForward for touch events.
2009         https://bugs.webkit.org/show_bug.cgi?id=48777
2010
2011         * DumpRenderTree/chromium/EventSender.cpp:
2012         (EventSender::sendCurrentTouchEvent):
2013
2014 2010-11-02  Adam Roben  <aroben@apple.com>
2015
2016         Only track resource identifiers in DRT when dumpResourceLoadCallbacks
2017         is on
2018
2019         This reverts Windows to our pre-r71097 behavior. That patch made us
2020         track all resource identifiers, including the main resource, so the
2021         main resource's URL started appearing in test output instead of
2022         "<unknown>". Arguably having the main resource's URL is better, but all
2023         other platforms print "<unknown>" and we want to match.
2024
2025         Fixes <http://webkit.org/b/48837> <rdar://problem/8620351> REGRESSION
2026         (r71097): Many http tests failing on Windows
2027
2028         Reviewed by Anders Carlsson.
2029
2030         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2031         (ResourceLoadDelegate::identifierForInitialRequest): Don't add the
2032         identifier to the URL map if we're not supposed to dump resource load
2033         callbacks.
2034         (ResourceLoadDelegate::removeIdentifierForRequest): Always remove the
2035         identifier from the URL map even if we're already "done". There's no
2036         point in keeping out-of-date identifiers around.
2037
2038 2010-11-01  Jenn Braithwaite  <jennb@chromium.org>
2039
2040         Reviewed by Adam Roben.
2041
2042         Windows: Update resource tracking when moving a frame between documents
2043         https://bugs.webkit.org/show_bug.cgi?id=48364
2044
2045         * DumpRenderTree/win/DumpRenderTree.cpp:
2046         (createWebViewAndOffscreenWindow):
2047         (main):
2048         Give each WebView its own ResourceLoadDelegate instance in order to
2049         make assertions about resource ids on a particular WebView.
2050         * DumpRenderTree/win/ResourceLoadDelegate.cpp:
2051         (ResourceLoadDelegate::identifierForInitialRequest):
2052         Always add id to the map.
2053         (ResourceLoadDelegate::removeIdentifierForRequest):
2054         Added.
2055         (ResourceLoadDelegate::willSendRequest):
2056         (ResourceLoadDelegate::didReceiveAuthenticationChallenge):
2057         (ResourceLoadDelegate::didReceiveResponse):
2058         (ResourceLoadDelegate::didFinishLoadingFromDataSource):
2059         (ResourceLoadDelegate::didFailLoadingWithError):
2060         (ResourceLoadDelegate::descriptionSuitableForTestResult):
2061         Replace static descriptionSuitableForTestResult with static member function to access identifier map.
2062         * DumpRenderTree/win/ResourceLoadDelegate.h:
2063         (ResourceLoadDelegate::urlMap):
2064         Moved within class so that each WebView has its own id map.
2065
2066 2010-11-01  Ojan Vafai  <ojan@chromium.org>
2067
2068         Reviewed by Dimitri Glazkov.
2069
2070         remove debug code from run_webkit_tests.py
2071         https://bugs.webkit.org/show_bug.cgi?id=48800
2072
2073         Remove temporary debug code and make --master-name required
2074         if --test-results-server is set now that all clients set
2075         --master-name.
2076
2077         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2078
2079 2010-11-01  Tony Chang  <tony@chromium.org>
2080
2081         Reviewed by Kent Tamura.
2082
2083         add plugin event logging for linux
2084         https://bugs.webkit.org/show_bug.cgi?id=48779
2085
2086         This is taken from Chromium's fork of the layout test plugin:
2087         http://git.chromium.org/gitweb/?p=chromium.git/.git;a=blob;f=webkit/tools/npapi_layout_test_plugin/main.cpp;h=3ebdada2f049b3624756438cff852364f86a2ede;hb=HEAD#l348
2088
2089         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2090         (handleEventX11):
2091         (NPP_HandleEvent):
2092
2093 2010-11-01  Dirk Pranke  <dpranke@chromium.org>
2094
2095         Reviewed by James Robinson.
2096
2097         new-run-webkit-tests: use DRT, child-processes=1 for GPU tests by default
2098         https://bugs.webkit.org/show_bug.cgi?id=48790
2099
2100         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2101         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2102
2103 2010-11-01  Adam Barth  <abarth@webkit.org>
2104
2105         Reviewed by Eric Seidel.
2106
2107         Teach check-webkit-style how to accept a list of files to diff on the
2108         command line
2109         https://bugs.webkit.org/show_bug.cgi?id=48784
2110
2111         In a future patch, webkit-patch will use this option to improve
2112         performance.  I'm landing this in two pieces to avoid causing a version
2113         skew problem for the style-bot.
2114
2115         * Scripts/check-webkit-style:
2116         * Scripts/webkitpy/style/optparser.py:
2117         * Scripts/webkitpy/style/optparser_unittest.py:
2118         * Scripts/webkitpy/style_references.py:
2119
2120 2010-11-01  Anders Carlsson  <andersca@apple.com>
2121
2122         Reviewed by John Sullivan.
2123
2124         Tear down the related WebProcessProxy when a WebContext is deallocated
2125         https://bugs.webkit.org/show_bug.cgi?id=48769
2126
2127         * TestWebKitAPI/Tests/WebKit2/FailedLoad.cpp:
2128         (TestWebKitAPI::didFailProvisionalLoadWithErrorForFrame):
2129         We don't support empty URLs anymore, update test to expect a null URL instead.
2130
2131 2010-11-01  Søren Gjesse  <sgjesse@chromium.org>
2132
2133         Reviewed by Andreas Kling.
2134
2135         Fix warning when compiling the chromium port of DumpRenderShell
2136         with clang.
2137         https://bugs.webkit.org/show_bug.cgi?id=48414
2138
2139         * DumpRenderTree/chromium/TestShell.h:
2140         (TestShell::javaScriptFlagsForLoad):
2141
2142 2010-11-01  Adam Roben  <aroben@apple.com>
2143
2144         Fix typo from r71022
2145
2146         * BuildSlaveSupport/build.webkit.org-config/config.json:
2147
2148 2010-11-01  Adam Roben  <aroben@apple.com>
2149
2150         Trigger the Windows Release WebKit2 tests when a Release build
2151         finishes, not when a Debug build finishes
2152
2153         Fixes <http://webkit.org/b/48754> Windows Release WebKit2 tests are
2154         triggered at the wrong time
2155
2156         Reviewed by Sam Weinig.
2157
2158         * BuildSlaveSupport/build.webkit.org-config/config.json: Fixed
2159         triggerable name and trigger.
2160
2161 2010-11-01  Mario Sanchez Prada  <msanchez@igalia.com>
2162
2163         Unreviewed. Adding my IRC nickname to the list of committers.
2164
2165         * Scripts/webkitpy/common/config/committers.py:
2166
2167 2010-10-31  Robert Hogan  <robert@webkit.org>
2168
2169         Reviewed by Antonio Gomes.
2170
2171         [Qt] [Gtk] Plug-ins having upper case in mime type are failing to load
2172
2173         Qt and Gtk are case-sensitive when storing the declared mime-type
2174         of plugins. Since plugin mime-types are lowercased prior to searching
2175         for them in the plugin database, ensure they are loaded with the
2176         mime-type in lower case too.
2177
2178         Change the test netscape plugin to declare its mimetype in sentence
2179         case so that the correct behaviour is enforced.
2180
2181         https://bugs.webkit.org/show_bug.cgi?id=36815
2182
2183         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2184         (NP_GetMIMEDescription):
2185
2186 2010-10-31  Robert Hogan  <robert@webkit.org>
2187
2188         Reviewed by Antonio Gomes.
2189
2190         [Qt] Support nodesFromRect in DRT
2191
2192         https://bugs.webkit.org/show_bug.cgi?id=48716
2193
2194         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2195         (LayoutTestController::nodesFromRect):
2196         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2197
2198 2010-10-30  Andreas Kling  <kling@webkit.org>
2199
2200         Reviewed by Antonio Gomes.
2201
2202         [Qt] QtTestBrowser: Switching view type moves the embedded inspector
2203         https://bugs.webkit.org/show_bug.cgi?id=48705
2204
2205         Reinsert the embedded inspector into the splitter after changing
2206         between QWebView/QGraphicsWebView.
2207
2208         * QtTestBrowser/launcherwindow.cpp:
2209         (LauncherWindow::init):
2210         (LauncherWindow::initializeView):
2211         * QtTestBrowser/webinspector.h:
2212         (WebInspector::WebInspector):
2213
2214 2010-10-28  Antonio Gomes  <agomes@rim.com>
2215
2216         Reviewed by Ojan Vafai.
2217
2218         Needs a "LinuxEditingBehavior", perhaps with a better name
2219         https://bugs.webkit.org/show_bug.cgi?id=36627
2220
2221         Adding support to Mac's, GTK+'s, Windows' and Chromium's LayoutTestController class to test the newly introduced Unix editing behavior.
2222
2223         * DumpRenderTree/chromium/LayoutTestController.cpp:
2224         (LayoutTestController::setEditingBehavior):
2225         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2226         (LayoutTestController::setEditingBehavior):
2227         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2228         (LayoutTestController::setEditingBehavior):
2229         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2230         (LayoutTestController::setEditingBehavior):
2231
2232 2010-10-29  Andreas Kling  <kling@webkit.org>
2233
2234         Reviewed by Kenneth Rohde Christiansen.
2235
2236         [Qt] QtTestBrowser: Add keyboard shortcut to toggle full screen (F11)
2237         https://bugs.webkit.org/show_bug.cgi?id=48695
2238
2239         * QtTestBrowser/launcherwindow.cpp:
2240         (LauncherWindow::createChrome):
2241
2242 2010-10-29  Andreas Kling  <kling@webkit.org>
2243
2244         Reviewed by Kenneth Rohde Christiansen.
2245
2246         [Qt] QtTestBrowser: Fix uninitialized read in FpsTimer
2247         https://bugs.webkit.org/show_bug.cgi?id=48675
2248
2249         FpsTimer::m_timer was never initialized and passed to QObject::killTimer()
2250         on startup with in -graphicsbased mode.
2251
2252         * QtTestBrowser/fpstimer.cpp:
2253         (FpsTimer::FpsTimer):
2254         (FpsTimer::numFrames):
2255         (FpsTimer::stop):
2256         (FpsTimer::timerEvent):
2257         * QtTestBrowser/fpstimer.h:
2258
2259 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2260
2261         Reviewed by Tony Chang.
2262
2263         new-run-webkit-tests: change TestResults to be serializable
2264
2265         In preparation for changing new-run-webkit-tests from
2266         multithreaded to multiprocess, we need to make sure the data
2267         going between the threads is easily serialized over a socket.
2268
2269         This change adds serialization/pickling for the TestResults and
2270         TestFailure objects (using cPickle).
2271
2272         The TestFailure objects included a "has_wdiff" flag for Text
2273         results, but the flag wasn't being used, so I've removed it,
2274         simplifying the state to basically a set of enum objects with
2275         associated methods.
2276
2277         https://bugs.webkit.org/show_bug.cgi?id=48616
2278
2279         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2280         * Scripts/webkitpy/layout_tests/layout_package/printing_unittest.py:
2281         * Scripts/webkitpy/layout_tests/layout_package/test_failures.py:
2282         * Scripts/webkitpy/layout_tests/layout_package/test_failures_unittest.py:
2283         * Scripts/webkitpy/layout_tests/layout_package/test_results.py: Added.
2284         * Scripts/webkitpy/layout_tests/layout_package/test_results_unittest.py:Added. 
2285         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2286         * Scripts/webkitpy/layout_tests/test_types/text_diff.py:
2287
2288 2010-10-29  Dirk Pranke  <dpranke@chromium.org>
2289
2290         Reviewed by Ojan Vafai.
2291
2292         new-run-webkit-tests: need to provide separate dashboard results for GPU tests
2293         https://bugs.webkit.org/show_bug.cgi?id=48687
2294
2295         Modify the steps to upload the JSON files to the dashboards so
2296         that the GPU tests don't conflict with the regular tests on a
2297         port. We do this by modifying the --builder-name parameter to
2298         append " - GPU", which should cause the app to treat the results
2299         as a completely new builder. This is a little non-obvious, but
2300         keeps us from having to restructure the app.
2301
2302         * Scripts/webkitpy/layout_tests/port/chromium_gpu.py:
2303         * Scripts/webkitpy/layout_tests/port/chromium_gpu_unittest.py:
2304         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2305
2306 2010-10-29  Ojan Vafai  <ojan@chromium.org>
2307
2308         Reviewed by Tony Chang.
2309
2310         [chromium] add debug logging to help diagnose flakiness dashboard issues
2311         https://bugs.webkit.org/show_bug.cgi?id=48657
2312
2313         The appengine app thinks it's getting empty files uploaded. Add some logging
2314         to see if new-run-webkit-tests agrees.
2315         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2316
2317 2010-10-29  Dimitri Glazkov  <dglazkov@chromium.org>
2318
2319         Reviewed by Ojan Vafai.
2320
2321         [Chromium/DRT] Add master-name flag to new-run-webkit-tests.
2322         https://bugs.webkit.org/show_bug.cgi?id=48649
2323
2324         * BuildSlaveSupport/build.webkit.org-config/master.cfg: Added flag.
2325
2326 2010-10-29  Nicolas Dufresne  <nicolas.dufresne@collabora.co.uk>
2327
2328         Reviewed by Martin Robinson.
2329
2330         Enable popup window in GtkLauncher
2331         https://bugs.webkit.org/show_bug.cgi?id=48335
2332
2333         The GtkLauncher application does not support opening new window when
2334         a link with "target=_blank" is clicked or similar call to
2335         window.open(). Instead, GtkLauncher does nothing which breaks
2336         navigation of some websites.
2337
2338         * GtkLauncher/main.c:
2339         (activate_uri_entry_cb):
2340         (update_title):
2341         (link_hover_cb):
2342         (notify_title_cb):
2343         (notify_load_status_cb):
2344         (notify_progress_cb):
2345         (destroy_cb):
2346         (go_back_cb):
2347         (go_forward_cb):
2348         (create_web_view_cb):
2349         (web_view_ready_cb):
2350         (close_web_view_cb):
2351         (create_browser):
2352         (create_statusbar):
2353         (create_toolbar):
2354         (create_window):
2355         (main):
2356
2357 2010-10-29  Adam Roben  <aroben@apple.com>
2358
2359         Teach check-webkit-style about WebKit2's idiosyncracies
2360
2361         Fixes <http://webkit.org/b/48638> Style bot complains about a number
2362         of WebKit2 conventions
2363
2364         Reviewed by Anders Carlsson.
2365
2366         * Scripts/webkitpy/style/checker.py: Excluded some rules for various
2367         WebKit2-related files. Also updated the excluded rules for
2368         WebKitAPITest to match the current code.
2369
2370 2010-10-29  Csaba Osztrogonác  <ossy@webkit.org>
2371
2372         Reviewed by Adam Roben and David Kilzer.
2373
2374         Fix and cleanup of build systems
2375         https://bugs.webkit.org/show_bug.cgi?id=48342
2376
2377         * Scripts/build-webkit:
2378          - Remove unnecessary ENABLE_SANDBOX option.
2379          - Add ENABLE_FULLSCREEN_API option.
2380
2381 2010-10-28  Adam Roben  <aroben@apple.com>
2382
2383         Switch the Windows WebKit2 bot to the Release configuration
2384
2385         We only have one machine testing WebKit2 on Windows right now, and
2386         Debug is just too slow for it to keep up.
2387
2388         Fixes (hopefully!) <http://webkit.org/b/48615> Windows WebKit2 bot is
2389         always way behind
2390
2391         Reviewed by Jon Honeycutt.
2392
2393         * BuildSlaveSupport/build.webkit.org-config/config.json:
2394
2395 2010-10-29  Leandro Gracia Gil  <leandrogracia@google.com>
2396
2397         Reviewed by Jeremy Orlow.
2398
2399         Added a second parameter to setMockSpeechInputResult for
2400         the language used in speech input.
2401         https://bugs.webkit.org/show_bug.cgi?id=47089
2402
2403         * DumpRenderTree/LayoutTestController.cpp:
2404         (setMockSpeechInputResultCallback):
2405         * DumpRenderTree/LayoutTestController.h:
2406         * DumpRenderTree/chromium/LayoutTestController.cpp:
2407         (LayoutTestController::setMockSpeechInputResult):
2408         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2409         (LayoutTestController::setMockSpeechInputResult):
2410         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2411         (LayoutTestController::setMockSpeechInputResult):
2412         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2413         (LayoutTestController::setMockSpeechInputResult):
2414         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2415         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2416         (LayoutTestController::setMockSpeechInputResult):
2417         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2418         (LayoutTestController::setMockSpeechInputResult):
2419         (LayoutTestController::pageSizeAndMarginsInPixels):
2420
2421 2010-10-28  Ojan Vafai  <ojan@chromium.org>
2422
2423         Reviewed by Tony Chang.
2424
2425         [chromium] add a result-small.json file for the test dashboard
2426         https://bugs.webkit.org/show_bug.cgi?id=48547
2427
2428         Output both a results.json file and a results-small.json file.
2429         The dashboard will load results-small.json by default so it loads faster.
2430
2431         * TestResultServer/model/jsonresults.py:
2432         * TestResultServer/model/jsonresults_unittest.py:
2433         Added a bunch of sys.path hackery. Unforunately, this uses hardcoded
2434         paths. That obviously needs to be fixed, but at least this way it
2435         clearly documents what paths are necessary.
2436
2437 2010-10-28  Eric Seidel  <eric@webkit.org>
2438
2439         Reviewed by Adam Barth.
2440
2441         webkit-patch upload calls changed_files more often than it should
2442         https://bugs.webkit.org/show_bug.cgi?id=48567
2443
2444         Passing changed_files around everywhere isn't a very elegant solution
2445         but it's the one we have for the moment.  I think keeping an explicit
2446         cache on Checkout (or making StepState() a real class) is a better
2447         long-term option.
2448
2449         Previously bug_id_for_this_commit was calling changed_files and the
2450         result was never getting cached on the state.  Now we're explicitly
2451         caching the result on the state and passing that to the bug_id_for_this_commit call.
2452
2453         I looked into building unit tests for this.  Doing so would require
2454         using a real Checkout object with a MockSCM and overriding the appropriate
2455         calls on SCM to count how often we're stating the file system.
2456         That's a useful set of tests to build for a separate change.
2457
2458         * Scripts/webkitpy/common/checkout/api.py:
2459         * Scripts/webkitpy/tool/commands/download.py:
2460         * Scripts/webkitpy/tool/commands/upload.py:
2461         * Scripts/webkitpy/tool/mocktool.py:
2462
2463 2010-10-28  Eric Seidel  <eric@webkit.org>
2464
2465         Reviewed by Adam Barth.
2466
2467         Make suggest-reviewers slightly faster
2468         https://bugs.webkit.org/show_bug.cgi?id=48562
2469
2470         Add @memoized to one more common call.
2471
2472         * Scripts/webkitpy/common/checkout/api.py:
2473
2474 2010-10-28  Eric Seidel  <eric@webkit.org>
2475
2476         Reviewed by Adam Barth.
2477
2478         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is still very slow
2479         https://bugs.webkit.org/show_bug.cgi?id=48536
2480
2481         This patch makes it about 40 seconds faster, but it still
2482         takes 1:40.  This will require more refinement.
2483
2484         The suggested reviewers list appears to be the same.
2485
2486         I think the next step may be to have it stop the search after
2487         5 reviewers are found.  We never want to suggest 30 people.
2488
2489         * Scripts/webkitpy/common/checkout/scm.py:
2490          - Using --remove-empty to theoretically stop lookups past
2491            when a file is removed.  I'm not entirely clear that the option
2492            does what it says it does.  Example:
2493            git log --pretty=format:%H -5 --remove-empty -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2494            returns only one commit
2495            vs.
2496            git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2497            which returns 5.  I was not aware that wx files were ever removed from the repository?
2498
2499 2010-10-28  Kinuko Yasuda  <kinuko@chromium.org>
2500
2501         Reviewed by David Levin.
2502
2503         [Chromium] Support FileSystem in chromium DRT
2504         https://bugs.webkit.org/show_bug.cgi?id=47643
2505
2506         * DumpRenderTree/chromium/WebViewHost.cpp:
2507         (WebViewHost::openFileSystem): Added.
2508         * DumpRenderTree/chromium/WebViewHost.h:
2509         (WebViewHost::openFileSystem): Added.
2510
2511 2010-10-28  Sam Weinig  <sam@webkit.org>
2512
2513         Reviewed by Anders Carlsson.
2514
2515         WKURLRefs should be allowed to be null
2516         <rdar://problem/8575621>
2517         https://bugs.webkit.org/show_bug.cgi?id=48535
2518
2519         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj:
2520         * TestWebKitAPI/Tests/WebKit2/PageLoadBasic.cpp:
2521         (TestWebKitAPI::didStartProvisionalLoadForFrame):
2522         (TestWebKitAPI::didCommitLoadForFrame):
2523         (TestWebKitAPI::didFinishLoadForFrame):
2524         Test that URLs are null pointers when unset.
2525
2526 2010-10-28  Chris Fleizach  <cfleizach@apple.com>
2527
2528         Reviewed by Adele Peterson.
2529
2530         AX: multi select group option does not handle setting of AXSelectedChildren correctly
2531         https://bugs.webkit.org/show_bug.cgi?id=48464
2532
2533         Add support for querying information about selected children to DRT, including:
2534            selectedChildAtIndex
2535            selectedChildrenCount
2536            setSelectedChild
2537
2538         * DumpRenderTree/AccessibilityUIElement.cpp:
2539         (selectedChildAtIndexCallback):
2540         (setSelectedChildCallback):
2541         (selectedChildrenCountCallback):
2542         (AccessibilityUIElement::setSelectedChild):
2543         (AccessibilityUIElement::selectedChildrenCount):
2544         (AccessibilityUIElement::selectedChildAtIndex):
2545         (AccessibilityUIElement::getJSClass):
2546         * DumpRenderTree/AccessibilityUIElement.h:
2547         * DumpRenderTree/mac/AccessibilityUIElementMac.mm:
2548         (AccessibilityUIElement::selectedChildAtIndex):
2549         (AccessibilityUIElement::selectedChildrenCount):
2550         (AccessibilityUIElement::setSelectedChild):
2551
2552 2010-10-27  Sam Weinig  <sam@webkit.org>
2553
2554         Reviewed by Anders Carlsson.
2555
2556         Add WebKit2 API for window feature getter/setters
2557         <rdar://problem/8590373>
2558         https://bugs.webkit.org/show_bug.cgi?id=48496
2559
2560         * MiniBrowser/mac/BrowserWindowController.m:
2561         (-[BrowserWindowController awakeFromNib]):
2562         * MiniBrowser/win/BrowserView.cpp:
2563         (BrowserView::create):
2564         * WebKitTestRunner/TestController.cpp:
2565         (WTR::createOtherPage):
2566         (WTR::TestController::initialize):
2567
2568 2010-10-28  Søren Gjesse  <sgjesse@chromium.org>
2569
2570         Reviewed by Tony Chang.
2571
2572         Added support for the DumpRenderTree flags --multiple-loads and --js-flags to the Python test runner.
2573         https://bugs.webkit.org/show_bug.cgi?id=48236
2574
2575         * Scripts/webkitpy/layout_tests/port/chromium.py:
2576         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2577
2578 2010-10-28  Sergio Villar Senin  <svillar@igalia.com>
2579
2580         Reviewed by Martin Robinson.
2581
2582         [GTK] WebKitWebFrame's load-status is not properly notified to the tests
2583         https://bugs.webkit.org/show_bug.cgi?id=48048
2584
2585         DRT now listens to WebKitWebFrame load-status signals for each
2586         created frame instead of the load-status signal of the
2587         WebKitWebView that only notifies about changes in the main frame.
2588
2589         * DumpRenderTree/gtk/DumpRenderTree.cpp:
2590         (webViewLoadFinished):
2591         (webFrameLoadStatusNotified):
2592         (frameCreatedCallback):
2593         (createWebView):
2594         (main):
2595
2596 2010-10-28  Tony Chang  <tony@chromium.org>
2597
2598         Reviewed by Kent Tamura.
2599
2600         [chromium] fix textInputController.{selectedRange,markedRange}
2601         https://bugs.webkit.org/show_bug.cgi?id=48487
2602
2603         * DumpRenderTree/chromium/TextInputController.cpp:
2604         (TextInputController::markedRange): Return arrays of ints, rather than a string
2605         (TextInputController::selectedRange): Ditto.
2606
2607 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2608
2609         Reviewed by Tony Chang.
2610
2611         [chromium] add a master-name flag to new-run-webkit-tests
2612         https://bugs.webkit.org/show_bug.cgi?id=48488
2613
2614         The test results server now allows adding a master name to
2615         the uploaded files. This lets us distinguish bots that have
2616         the same name, but are on different masters.
2617
2618         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2619
2620 2010-10-28  Adam Roben  <aroben@apple.com>
2621
2622         Don't append a newline to the test output if the frame has no document
2623         element in WebKitTestRunner
2624
2625         Fixes <http://webkit.org/b/48526> Extra trailing newline when running
2626         plugins/document-open.html in WebKitTestRunner
2627
2628         Reviewed by Anders Carlsson.
2629
2630         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2631         (WTR::toJS): Added. Turns a UTF-8 C string into a JSStringRef.
2632         (WTR::hasDocumentElement): Added. Uses the JSC API to figure out
2633         whether the frame has a document element.
2634         (WTR::dumpFrameText): Match DRT by bailing (rather than appending an
2635         empty string and a newline) if the frame has no document element.
2636
2637 2010-10-28  Adam Roben  <aroben@apple.com>
2638
2639         Skip npn-invalidate-rect-invalidates-window.html on headless XP
2640         machines
2641
2642         TestNetscapePlugIn never receives a WM_PAINT message on headless XP
2643         machines, so this test times out. Fixing the test is covered by
2644         <http://webkit.org/b/48333>.
2645
2646         * Scripts/old-run-webkit-tests: Skip
2647         npn-invalidate-rect-invalidates-window.html on Windows if accelerated
2648         compositing support is disabled, which likely means we're on a headless
2649         XP machine.
2650
2651 2010-10-28  Kimmo Kinnunen  <kimmok@iki.fi>
2652
2653         Adding myself as a committer.
2654
2655         * Scripts/webkitpy/common/config/committers.py:
2656
2657 2010-10-27  Eric Seidel  <eric@webkit.org>
2658
2659         Reviewed by Adam Barth.
2660
2661         webkit-patch suggest-reviewers -g 260550a6e30b7bf34f16bdb4a5396cf26264fc1c is very slow
2662         https://bugs.webkit.org/show_bug.cgi?id=48500
2663
2664         This doesn't fix the problem, but it makes things slightly better.
2665         Each git svn find-rev call takes about .25 seconds on my desktop.
2666         This patch uses a new memoized class to avoid those calls when possible.
2667
2668         The real slowness is still git log on some files, like:
2669         git log --pretty=format:%H -5 -- /Projects/WebKit/WebCore/platform/wx/SearchPopupMenuWx.h
2670         I'm not yet sure how to make the pathological git logs better.
2671
2672         * Scripts/webkitpy/common/checkout/scm.py:
2673         * Scripts/webkitpy/common/memoized.py: Added.
2674         * Scripts/webkitpy/common/memoized_unittest.py: Added.
2675
2676 2010-10-27  Eric Seidel  <eric@webkit.org>
2677
2678         Reviewed by Ojan Vafai.
2679
2680         EWS bots should not use --quiet when running build-webkit
2681         https://bugs.webkit.org/show_bug.cgi?id=48482
2682
2683         --quiet is only correct when the sub-process does the error reporting.
2684         In the case of _can_build() the parent process is reporting the error.
2685         We'd like the full build log at queues.webkit.org so someone can look
2686         at the log and understand why the EWS is failing to build trunk.
2687
2688         * Scripts/webkitpy/layout_tests/port/chromium_win_unittest.py:
2689         * Scripts/webkitpy/tool/commands/earlywarningsystem.py:
2690         * Scripts/webkitpy/tool/commands/earlywarningsystem_unittest.py:
2691         * Scripts/webkitpy/tool/mocktool.py:
2692
2693 2010-10-27  Brian Weinstein  <bweinstein@apple.com>
2694
2695         More Windows build fixage. Rename a variable that was named string.
2696
2697         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2698         (WTR::hasPrefix):
2699
2700 2010-10-27  Ojan Vafai  <ojan@chromium.org>
2701
2702         Reviewed by Dimitri Glazkov.
2703
2704         [chromium] Make the test results server store which master the bot is on
2705         https://bugs.webkit.org/show_bug.cgi?id=48478
2706
2707         The chromium bots recently changed so that there are multiple slaves with
2708         the same name on different masters. Up till now, the test results server
2709         assumed slave names were unique. Adds a master field to the file in order
2710         to distinguish.
2711
2712         Also, for files that currently lack a master or testtype, set them appropriately.
2713
2714         * TestResultServer/handlers/testfilehandler.py:
2715         * TestResultServer/index.yaml:
2716         * TestResultServer/model/jsonresults.py:
2717         * TestResultServer/model/testfile.py:
2718         * TestResultServer/templates/showfilelist.html:
2719         * TestResultServer/templates/uploadform.html:
2720
2721 2010-10-26  Darin Adler  <darin@apple.com>
2722
2723         Reviewed by Sam Weinig.
2724
2725         WebKitTestRunner needs to support layoutTestController.dumpBackForwardList
2726         https://bugs.webkit.org/show_bug.cgi?id=42322
2727         rdar://problem/8193631
2728
2729         WebKitTestRunner needs to support layoutTestController.clearBackForwardList
2730         https://bugs.webkit.org/show_bug.cgi?id=42333
2731         rdar://problem/8193643
2732
2733         * WebKitTestRunner/Configurations/InjectedBundle.xcconfig:
2734         Renamed the product to WebKitTestRunnerInjectedBundle to avoid
2735         name conflicts in the build directory.
2736
2737         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
2738         Added dumpBackForwardList and clearBackForwardList.
2739
2740         * WebKitTestRunner/InjectedBundle/InjectedBundle.cpp:
2741         (WTR::InjectedBundle::InjectedBundle): Removed initialization of m_mainPage.
2742         (WTR::InjectedBundle::didCreatePage): Removed unneeded underscore.
2743         (WTR::InjectedBundle::willDestroyPage): Ditto.
2744         (WTR::InjectedBundle::didReceiveMessage): Ditto.
2745         (WTR::InjectedBundle::initialize): Ditto.
2746         (WTR::InjectedBundle::didCreatePage): Changed code to use m_pages
2747         instead of m_mainPage and m_otherPages.
2748         (WTR::InjectedBundle::willDestroyPage): Ditto.
2749         (WTR::InjectedBundle::page): Ditto.
2750         (WTR::InjectedBundle::beginTesting): Ditto.
2751         (WTR::InjectedBundle::done): Ditto.
2752         (WTR::InjectedBundle::closeOtherPages): Ditto.
2753         (WTR::InjectedBundle::dumpBackForwardListsForAllPages): Added.
2754
2755         * WebKitTestRunner/InjectedBundle/InjectedBundle.h: Changed
2756         page function to longer be inline, and pageCount function
2757         to use m_pages. Added dumpBackForwardListsForAllPages, and
2758         removed some underscores. Replaced m_mainPage and m_otherPageas
2759         with m_pages.
2760
2761         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
2762         (WTR::adoptWK): Added.
2763         (WTR::hasPrefix): Added.
2764         (WTR::InjectedBundlePage::reset): Added code to set up
2765         m_previousTestBackForwardListItem.
2766         (WTR::InjectedBundlePage::dump): Added code to call
2767         dumpBackForwardListsForAllPages.
2768         (WTR::compareByTargetName): Added.
2769         (WTR::dumpBackForwardListItem): Added.
2770         (WTR::InjectedBundlePage::dumpBackForwardList): Added.
2771
2772         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h: Added
2773         dumpBackForwardList and m_previousTestBackForwardListItem.
2774
2775         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
2776         (WTR::LayoutTestController::LayoutTestController): Initialize
2777         m_shouldDumpBackForwardListsForAllWindows to false.
2778         (WTR::LayoutTestController::clearBackForwardList): Added.
2779
2780         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
2781         Added dumpBackForwardList, clearBackForwardList,
2782         shouldDumpBackForwardListsForAllWindows, and
2783         m_shouldDumpBackForwardListsForAllWindows.
2784
2785         * WebKitTestRunner/TestController.cpp:
2786         (WTR::TestController::initialize): Set up
2787         didReceiveSynchronousMessageFromInjectedBundle.
2788         (WTR::TestController::didReceiveSynchronousMessageFromInjectedBundle):
2789         Added.
2790         * WebKitTestRunner/TestController.h: Ditto.
2791
2792         * WebKitTestRunner/TestInvocation.cpp:
2793         (WTR::TestInvocation::didReceiveSynchronousMessageFromInjectedBundle):
2794         Added.
2795         * WebKitTestRunner/TestInvocation.h: Ditto.
2796
2797         * WebKitTestRunner/WebKitTestRunner.xcodeproj/project.pbxproj:
2798         Renamed bundle to WebKitTestRunnerInjectedBundle (see above).
2799         * WebKitTestRunner/mac/TestControllerMac.mm:
2800         (WTR::TestController::initializeInjectedBundlePath): Ditto.
2801
2802 2010-10-27  Chris Rogers  <crogers@google.com>
2803
2804         Reviewed by Chris Marrin.
2805
2806         Add ENABLE_WEB_AUDIO feature enable flag (initially disabled) to build-webkit
2807         https://bugs.webkit.org/show_bug.cgi?id=48279
2808
2809         * Scripts/build-webkit:
2810
2811 2010-10-27  Eric Seidel  <eric@webkit.org>
2812
2813         Unreviewed.
2814
2815         build-webkit should collect Visual Studio Express logs and display them
2816         https://bugs.webkit.org/show_bug.cgi?id=39199
2817
2818         It turns out my previous patch wasn't actually working on the win-ews
2819         machine.  So I've fixed my mistakes from before.
2820
2821         * Scripts/build-webkit:
2822          - Windows VSE builds change the CWD while building.  Why?  Who knows.
2823         * Scripts/print-vse-failure-logs:
2824          - windows VSE builds don't use Debug/Release as I expected, so I've
2825            moved off of --configuration to --top-level and added the /obj
2826            optimization while I was there.
2827
2828 2010-10-27  Dimitri Glazkov  <dglazkov@chromium.org>
2829
2830         Unreviewed, rolling out r70674.
2831         http://trac.webkit.org/changeset/70674
2832         https://bugs.webkit.org/show_bug.cgi?id=48053
2833
2834         Broke Chromium Windows build.
2835
2836         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2837         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2838
2839 2010-10-25  Tony Chang  <tony@chromium.org>
2840
2841         Reviewed by Anders Carlsson.
2842
2843         compile TestNetscapePlugIn on chromium linux
2844         https://bugs.webkit.org/show_bug.cgi?id=48274
2845
2846         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2847         (NP_Initialize): On Linux, plugin funcs are set in initialize.
2848         (NPP_New): Mark the plugin as windowless.
2849         (NPP_GetValue): Handle mime type values.
2850         (NP_GetMIMEDescription):
2851         (NP_GetValue):
2852         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npapi.h: define TRUE and FALSE, which are in webkit's npapi.h.
2853         * DumpRenderTree/chromium/TestNetscapePlugIn/ForwardingHeaders/WebKit/npfunctions.h: Pull in npapi.h to get TRUE/FALSE (matches webkit's npfunctions.h)
2854
2855 2010-10-27  Kenneth Rohde Christiansen  <kenneth@webkit.org>
2856
2857         Reviewed by Anders Carlsson.
2858
2859         Remove contentSizeChanged callbacks as it is no longer
2860         part of the public UIClient.
2861
2862         Make WKPageContentsSizeChangedCallback be a private API
2863         https://bugs.webkit.org/show_bug.cgi?id=48409
2864
2865         * MiniBrowser/mac/BrowserWindowController.m:
2866         (-[BrowserWindowController awakeFromNib]):
2867         * MiniBrowser/win/BrowserView.cpp:
2868         (BrowserView::create):
2869         * WebKitTestRunner/TestController.cpp:
2870         (WTR::createOtherPage):
2871         (WTR::TestController::initialize):
2872
2873 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2874
2875         Reviewed by Ojan Vafai.
2876
2877         Make http locking default in NRWT.
2878         https://bugs.webkit.org/show_bug.cgi?id=48053
2879
2880         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
2881         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
2882
2883 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
2884
2885         Reviewed by Tony Chang.
2886
2887         [NRWT] Fix http lock on Windows platform
2888         https://bugs.webkit.org/show_bug.cgi?id=48321
2889
2890         * Scripts/webkitpy/layout_tests/port/http_lock.py:
2891
2892 2010-10-27  Satish Sampath  <satish@chromium.org>
2893
2894         Unreviewed, rolling out r70665.
2895         http://trac.webkit.org/changeset/70665
2896         https://bugs.webkit.org/show_bug.cgi?id=47089
2897
2898         Need to address Alexey's review comments.
2899
2900         * DumpRenderTree/LayoutTestController.cpp:
2901         (setMockSpeechInputResultCallback):
2902         * DumpRenderTree/LayoutTestController.h:
2903         * DumpRenderTree/chromium/LayoutTestController.cpp:
2904         (LayoutTestController::setMockSpeechInputResult):
2905         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2906         (LayoutTestController::setMockSpeechInputResult):
2907         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2908         (LayoutTestController::setMockSpeechInputResult):
2909         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2910         (LayoutTestController::setMockSpeechInputResult):
2911         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2912         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2913         (LayoutTestController::setMockSpeechInputResult):
2914         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2915         (LayoutTestController::setMockSpeechInputResult):
2916
2917 2010-10-27  Leandro Gracia Gil  <leandrogracia@google.com>
2918
2919         Reviewed by Jeremy Orlow.
2920
2921         Added a second parameter to setMockSpeechInputResult for
2922         the language used in speech input.
2923         https://bugs.webkit.org/show_bug.cgi?id=47089
2924
2925         * DumpRenderTree/LayoutTestController.cpp:
2926         (setMockSpeechInputResultCallback):
2927         * DumpRenderTree/LayoutTestController.h:
2928         * DumpRenderTree/chromium/LayoutTestController.cpp:
2929         (LayoutTestController::setMockSpeechInputResult):
2930         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
2931         (LayoutTestController::setMockSpeechInputResult):
2932         * DumpRenderTree/mac/LayoutTestControllerMac.mm:
2933         (LayoutTestController::setMockSpeechInputResult):
2934         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
2935         (LayoutTestController::setMockSpeechInputResult):
2936         * DumpRenderTree/qt/LayoutTestControllerQt.h:
2937         * DumpRenderTree/win/LayoutTestControllerWin.cpp:
2938         (LayoutTestController::setMockSpeechInputResult):
2939         * DumpRenderTree/wx/LayoutTestControllerWx.cpp:
2940         (LayoutTestController::setMockSpeechInputResult):
2941         (LayoutTestController::pageSizeAndMarginsInPixels):
2942
2943 2010-10-27  Adam Roben  <aroben@apple.com>
2944
2945         Reset TestNetscapePlugIn's NPP_GetValue pointer when
2946         NullNPPGetValuePointer finishes running
2947
2948         Fixes <http://webkit.org/b/48435> REGRESSION (r70655): Many plugins
2949         tests are failing on Qt
2950
2951         Reviewed by Anders Carlsson.
2952
2953         * DumpRenderTree/TestNetscapePlugIn/PluginTest.cpp:
2954         (PluginTest::NPP_Destroy):
2955         * DumpRenderTree/TestNetscapePlugIn/PluginTest.h:
2956         Added a do-nothing NPP_Destroy implementation.
2957
2958         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp:
2959         (NullNPPGetValuePointer::NullNPPGetValuePointer): Save the original
2960         NPP_GetValue pointer in m_originalNPPGetValuePointer so we can restore
2961         it later.
2962         (NullNPPGetValuePointer::NPP_Destroy): Added. Restores the original
2963         NPP_GetValue pointer so it can be used in other tests.
2964
2965         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
2966         (NPP_Destroy):
2967         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2968         (webkit_test_plugin_destroy_instance):
2969         Call through to the PluginTest when NPP_Destroy is called.
2970
2971 2010-10-26  Brian Weinstein  <bweinstein@apple.com>
2972
2973         Reviewed by Adam Roben.
2974
2975         WebKit2 shouldn't try to send an empty user agent
2976         https://bugs.webkit.org/show_bug.cgi?id=48397
2977         
2978         Add a test that when we set our custom user agent to an empty string, we don't send
2979         an empty user agent.
2980
2981         * TestWebKitAPI/Tests/WebKit2/PreventEmptyUserAgent.cpp: Added.
2982         (TestWebKitAPI::didRunJavaScript): Make sure that the result of navigator.userAgent isn't empty.
2983         (TestWebKitAPI::TEST): Set our custom user agent to the empty string, and run navigator.userAgent.
2984         
2985         * TestWebKitAPI/TestWebKitAPI.xcodeproj/project.pbxproj: Add the new file.
2986         * TestWebKitAPI/win/TestWebKitAPI.vcproj: Ditto.
2987
2988 2010-10-27  Adam Roben  <aroben@apple.com>
2989
2990         Check in file I forgot in r70653
2991
2992         * DumpRenderTree/unix/TestNetscapePlugin/TestNetscapePlugin.cpp:
2993         (NP_Initialize): Save the NPPluginFuncs struct the browser passed to us
2994         so we can be naughty and modify it later.
2995
2996 2010-10-27  Adam Roben  <aroben@apple.com>
2997
2998         Test that WebKit doesn't crash if the plugin passes 0 for its
2999         NPP_GetValue pointer
3000
3001         Test for <http://webkit.org/b/48433> Crash in
3002         NetscapePlugin::shouldLoadSrcURL when using Shockwave Director 10.3 in
3003         WebKit2 on Windows
3004
3005         Reviewed by Eric Carlson.
3006
3007         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3008         * DumpRenderTree/TestNetscapePlugIn/win/TestNetscapePlugin.vcproj:
3009         * DumpRenderTree/qt/TestNetscapePlugin/TestNetscapePlugin.pro:
3010         * GNUmakefile.am:
3011         Added NullNPPGetValuePointer.cpp.
3012
3013         * DumpRenderTree/TestNetscapePlugIn/PluginObject.cpp:
3014         * DumpRenderTree/TestNetscapePlugIn/PluginObject.h:
3015         Set up a place to store the NPPluginFuncs struct the browser passed to us.
3016
3017         * DumpRenderTree/TestNetscapePlugIn/Tests/NullNPPGetValuePointer.cpp: Added.
3018         (NullNPPGetValuePointer::NullNPPGetValuePointer): Null out the
3019         NPP_GetValue pointer we passed to the browser to simulate a plugin that
3020         doesn't implement NPP_GetValue.
3021         (NullNPPGetValuePointer::NPP_GetValue): Print an error message. If this
3022         function is called, it means that WebKit has changed in a way that
3023         makes this test invalid.
3024
3025         * DumpRenderTree/TestNetscapePlugIn/main.cpp:
3026         (NP_GetEntryPoints): Save the NPPluginFuncs struct the browser passed
3027         to us so we can be naughty and modify it later.
3028
3029 2010-10-27  Ademar de Souza Reis Jr  <ademar.reis@openbossa.org>
3030
3031         Reviewed by Andreas Kling.
3032
3033         Remove references to ancient QGVLauncher and QtLauncher
3034         https://bugs.webkit.org/show_bug.cgi?id=48430
3035
3036         QtTestBrowser substitutes both and has checks enabled.
3037
3038         * Scripts/webkitpy/style/checker.py: remove references
3039         * Scripts/webkitpy/style/checker_unittest.py: ditto
3040
3041 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3042
3043         Reviewed by Ojan Vafai.
3044
3045         [NRWT] Don't use image hash when it's no need in single test mode.
3046         https://bugs.webkit.org/show_bug.cgi?id=48326
3047
3048         * Scripts/webkitpy/layout_tests/layout_package/dump_render_tree_thread.py:
3049
3050 2010-10-27  David Kilzer  <ddkilzer@apple.com>
3051
3052         Fix leak of CFMutableDictionaryRef in createXMLStringFromWebArchiveData()
3053
3054         Reviewed by Adam Roben.
3055
3056         Follow-up fix for: <https://bugs.webkit.org/show_bug.cgi?id=48278>
3057
3058         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3059         (createXMLStringFromWebArchiveData): Use RetainPtr<> to fix a
3060         leak introduced in r70613.
3061
3062 2010-10-27  Gabor Rapcsanyi  <rgabor@inf.u-szeged.hu>
3063
3064         Reviewed by Csaba Osztrogonác.
3065
3066         [NRWT] Add platform specific baseline search paths for Qt port
3067         https://bugs.webkit.org/show_bug.cgi?id=48428
3068
3069         * Scripts/webkitpy/layout_tests/port/qt.py:
3070
3071 2010-10-27  Adam Roben  <aroben@apple.com>
3072
3073         Catch exceptions when checking if we're inside a git working directory
3074
3075         Fixes <http://webkit.org/b/48420> REGRESSION (r70562): test-webkitpy
3076         fails on systems without git installed
3077
3078         Reviewed by Anders Carlsson.
3079
3080         * Scripts/webkitpy/common/net/credentials.py:
3081         (Credentials._credentials_from_git): Put the call to
3082         Git.in_working_directory inside the try/except since it, too, attempts
3083         to execute git and thus will throw on systems that don't have git
3084         installed.
3085
3086 2010-10-27  Nikolas Zimmermann  <nzimmermann@rim.com>
3087
3088         Reviewed by David Kilzer.
3089
3090         Convert DumpRenderTree webarchive code to CoreFoundation
3091         https://bugs.webkit.org/show_bug.cgi?id=48278
3092
3093         CFPropertyListCreateWithData and CFPropertyListCreateData are only available in 10.6+.
3094         Replace CFPropertyListCreateWithData by a combination of CFReadStreamCreateWithBytesNoCopy and CFPropertyListCreateFromStream.
3095         Replace CFPropertyListCreateData by CFPropertyListCreateXMLData.
3096
3097         These changes are wrapped in BUILDING_ON_LEOPARD, as the methods are deprecated on 10.6+.
3098
3099         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp:
3100         (createXMLStringFromWebArchiveData):
3101
3102 2010-10-26  Antonio Gomes  <agomes@rim.com>
3103
3104         Reviewed by Martin Robinson.
3105
3106         Remove the absolute path used to include DumpRenderTreeSupportGtk.h from LayoutTestController.h
3107
3108         It turns out this is rather unneeded since WebKitTools/GNUMakefile.am has WebKit/gtk/ in its include
3109         path, and then we can just do #include "WebCoreSupport/DumpRenderTreeSupportGtk.h"
3110
3111         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3112
3113 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3114
3115         Rename WebArchiveDumpSupport.mm to WebArchiveDumpSupport.cpp
3116
3117         Reviewed by Adam Roben.
3118
3119         Part 4 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3120
3121         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj:
3122         Updated for file moves and renames.
3123         * DumpRenderTree/cf/WebArchiveDumpSupport.cpp: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.mm.
3124         * DumpRenderTree/cf/WebArchiveDumpSupport.h: Renamed from DumpRenderTree/mac/WebArchiveDumpSupport.h.
3125
3126 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3127
3128         Convert WebArchiveDumpSupport.mm from NS objects to CF types
3129
3130         Reviewed by Adam Roben.
3131
3132         Part 3 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3133
3134         In order to share WebArchive code between the Mac and Windows
3135         ports, the code in WebArchiveDumpSupport.mm was converted from
3136         Cocoa to C++ using CoreFoundation (CF) types.
3137
3138         * DumpRenderTree/mac/DumpRenderTree.mm:
3139         (dump): Renamed serializeWebArchiveToXML() to
3140         createXMLStringFromWebArchiveData() and added HardAutorelease()
3141         to prevent leaks.
3142         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3143         (createXMLStringFromWebArchiveData): Renamed from
3144         serializeWebArchiveToXML().  Changed to use CF types.
3145         (createCFURLResponseFromResponseData): Renamed from
3146         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3147         types for its parameter and return type.
3148         * DumpRenderTree/mac/WebArchiveDumpSupport.mm: Replaced use of
3149         NS objects with CF types.  It will be renamed to *.cpp in a
3150         future commit.
3151         (convertMIMEType): Changed to use case-insensitive string
3152         comparisons.
3153         (convertWebResourceDataToString):
3154         (normalizeHTTPResponseHeaderFields):
3155         (normalizeWebResourceURL):
3156         (convertWebResourceResponseToDictionary):
3157         (compareResourceURLs):
3158         (createXMLStringFromWebArchiveData):
3159         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3160         (createCFURLResponseFromResponseData): Renamed from
3161         unarchiveNSURLResponseFromResponseData().  Changed to use CF
3162         types for its parameter and return type.
3163
3164 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3165
3166         Extract use of NSKeyedUnarchiver from WebArchiveDumpSupport.mm
3167
3168         Reviewed by Adam Roben.
3169
3170         Part 2 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3171
3172         There is no equivalent to NSKeyedUnarchiver in CoreFoundation,
3173         so extract it into a platform-specific source file.
3174
3175         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3176         (unarchiveNSURLResponseFromResponseData): Added declaration.
3177         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3178         (convertWebResourceResponseToDictionary): Extracted code to
3179         unarchiveNSURLResponseFromResponseData() in
3180         WebArchiveDumpSupportMac.mm.  Updated logic to return early if
3181         nil is returned from unarchiveNSURLResponseFromResponseData().
3182         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm:
3183         (unarchiveNSURLResponseFromResponseData): Added.  Extracted
3184         code from convertWebResourceResponseToDictionary() in
3185         WebArchiveDumpSupport.mm.
3186
3187 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3188
3189         Extract call to -[WebHTMLRepresentation supportedNonImageMIMETypes] from WebArchiveDumpSupport.mm
3190
3191         Reviewed by Adam Roben.
3192
3193         Part 1 of 4: <http://webkit.org/b/48278> Convert DumpRenderTree webarchive code to CoreFoundation
3194
3195         The call to -[WebHTMLRepresentation supportedNonImageMIMETypes]
3196         is not cross-platform between Mac and Windows, so extract it
3197         into a platform-specific source file.
3198
3199         * DumpRenderTree/DumpRenderTree.xcodeproj/project.pbxproj: Added
3200         WebArchiveDumpSupportMac.mm to the project.
3201         * DumpRenderTree/mac/WebArchiveDumpSupport.h:
3202         (supportedNonImageMIMETypes): Added declaration.
3203         * DumpRenderTree/mac/WebArchiveDumpSupport.mm:
3204         (convertWebResourceDataToString): Replaced call to
3205         -[WebHTMLRepresentation supportedNonImageMIMETypes] with
3206         supportedNonImageMIMETypes().
3207         * DumpRenderTree/mac/WebArchiveDumpSupportMac.mm: Added.
3208         (supportedNonImageMIMETypes): Added.  Extracted from
3209         WebArchiveDumpSupport.mm.
3210
3211 2010-10-26  Antonio Gomes  <agomes@rim.com>
3212
3213         Reviewed by Martin Robinson.
3214
3215         [GTK] Implement DumpRenderTreeSupportGtk (similarly to DumpRenderTreeSupportQt idea)
3216         https://bugs.webkit.org/show_bug.cgi?id=48199
3217
3218         Implements support to WebKitTabToLinksPreferenceKey through LayoutTestController::overridePreference.
3219         The corresponding DumpRenderTreeSupportGtk method is called in the DRT context only.
3220
3221         * DumpRenderTree/gtk/DumpRenderTree.cpp:
3222         (resetDefaultsToConsistentValues):
3223         (createWebView):
3224         * DumpRenderTree/gtk/LayoutTestControllerGtk.cpp:
3225         (LayoutTestController::overridePreference):
3226
3227 2010-10-26  Ariya Hidayat  <ariya@sencha.com>
3228
3229         Unreviewed, change the order of my emails for bugzilla autocompletion.
3230
3231         * Scripts/webkitpy/common/config/committers.py:
3232
3233 2010-10-26  Eric Seidel  <eric@webkit.org>
3234
3235         Reviewed by David Kilzer.
3236
3237         build-webkit should collect Visual Studio Express logs and display them
3238         https://bugs.webkit.org/show_bug.cgi?id=39199
3239
3240         * Scripts/build-webkit:
3241         * Scripts/print-vse-failure-logs: Added.
3242         * Scripts/webkitdirs.pm:
3243
3244 2010-10-26  David Kilzer  <ddkilzer@apple.com>
3245
3246         <http://webkit.org/b/48224> build-webkit: add support for --meter-tag switch
3247
3248         Reviewed by Csaba Osztrogonác.
3249
3250         * Scripts/build-webkit: Added support for --meter-tag switch.
3251         It should have been added with r60820.
3252
3253 2010-10-26  Ademar de Souza Reis Jr.  <ademar.reis@openbossa.org>
3254
3255         Reviewed by Dumitru Daniliuc.
3256
3257         check-webkit-style fails on operator+=, operator-=, ... methods
3258         https://bugs.webkit.org/show_bug.cgi?id=48258
3259
3260         * Scripts/webkitpy/style/checkers/cpp.py: Added exceptions
3261         * Scripts/webkitpy/style/checkers/cpp_unittest.py: Added unit-tests
3262
3263 2010-10-26  Ojan Vafai  <ojan@chromium.org>
3264
3265         Reviewed by Tony Chang.
3266
3267         remove DEFER support from new-run-webkit-tests
3268         https://bugs.webkit.org/show_bug.cgi?id=48387
3269
3270         DEFER was needed when we were trying to ship Chrome beta.
3271         Now it's just extra complication.
3272
3273         * Scripts/webkitpy/layout_tests/layout_package/json_layout_results_generator.py:
3274         * Scripts/webkitpy/layout_tests/layout_package/test_expectations.py:
3275         * Scripts/webkitpy/layout_tests/layout_package/test_expectations_unittest.py:
3276         * Scripts/webkitpy/layout_tests/port/chromium_unittest.py:
3277         * Scripts/webkitpy/layout_tests/run_webkit_tests.py:
3278         * Scripts/webkitpy/style/checkers/test_expectations_unittest.py:
3279
3280 2010-10-26  Eric Seidel  <eric@webkit.org>
3281
3282         Unreviewed.  Adding import with_statement to fix python 2.5.
3283
3284         Teach webkit-patch how to read credentials from the environment
3285         https://bugs.webkit.org/show_bug.cgi?id=48275
3286
3287         * Scripts/webkitpy/common/net/credentials_unittest.py:
3288          - import with_statement to fix python 2.5.
3289
3290 2010-10-26  Alexey Proskuryakov  <ap@apple.com>
3291
3292         Reviewed by Sam Weinig.
3293
3294         https://bugs.webkit.org/show_bug.cgi?id=48375
3295         <rdar://problem/8392724> Need delegate calls in PageLoaderClient to indicate if we have
3296         loaded insecure content
3297
3298         Updated for WebKit2 changes.
3299
3300         * MiniBrowser/mac/BrowserWindowController.m:
3301         (didDisplayInsecureContentForFrame):
3302         (didRunInsecureContentForFrame):
3303         (-[BrowserWindowController awakeFromNib]):
3304         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3305         (WTR::InjectedBundlePage::InjectedBundlePage):
3306         (WTR::InjectedBundlePage::didDisplayInsecureContentForFrame):
3307         (WTR::InjectedBundlePage::didRunInsecureContentForFrame):
3308         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.h:
3309         * WebKitTestRunner/TestController.cpp:
3310         (WTR::TestController::initialize):
3311
3312 2010-10-26  Adam Roben  <aroben@apple.com>
3313
3314         Clear up confusion between "3D rendering" and "accelerated compositing"
3315
3316         WebKit has no software-based 3D rendering implementation, so 3D
3317         rendering can only be enabled when accelerated compositing is. But DRT
3318         was falsely reporting that 3D rendering was available on machines that
3319         don't support accelerated compositing, leading to much confusion.
3320
3321         Reviewed by Darin Adler.
3322
3323         Fixes <http://webkit.org/b/48370> REGRESSION (r70540): Many
3324         transforms/3d tests are failing on the XP bots
3325
3326         * DumpRenderTree/win/DumpRenderTree.cpp:
3327         (main): When ENABLE(3D_RENDERING) is turned on, only report that 3D
3328         rendering is available when acclerated compositing is also available.
3329
3330         * Scripts/old-run-webkit-tests: Skip tests that use the
3331         -webkit-transform-3d media query when 3D rendering is disabled,
3332         rather than when accelerated compositing is disabled.
3333
3334 2010-10-26  Eric Seidel  <eric@webkit.org>
3335
3336         Reviewed by Ojan Vafai.
3337
3338         Remove rietveld code now that it's unused
3339         https://bugs.webkit.org/show_bug.cgi?id=48359
3340
3341         Was cool that we added this support, but now that it's
3342         not used it makes little sense to keep it around.
3343         We can always restore this code from SVN if we need it.
3344
3345         * Scripts/webkitpy/common/config/__init__.py:
3346         * Scripts/webkitpy/common/net/bugzilla.py:
3347         * Scripts/webkitpy/common/net/bugzilla_unittest.py:
3348         * Scripts/webkitpy/common/net/rietveld.py: Removed.
3349         * Scripts/webkitpy/common/net/rietveld_unittest.py: Removed.
3350         * Scripts/webkitpy/thirdparty/__init__.py:
3351         * Scripts/webkitpy/tool/commands/download.py:
3352         * Scripts/webkitpy/tool/commands/download_unittest.py:
3353         * Scripts/webkitpy/tool/commands/queues.py:
3354         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3355         * Scripts/webkitpy/tool/main.py:
3356         * Scripts/webkitpy/tool/mocktool.py:
3357         * Scripts/webkitpy/tool/steps/__init__.py:
3358         * Scripts/webkitpy/tool/steps/postcodereview.py: Removed.
3359
3360 2010-10-26  Anders Carlsson  <andersca@apple.com>
3361
3362         Reviewed by Sam Weinig.
3363
3364         pageDidScroll callback should be on the UI process client rather than (or in addition to) the web process client
3365         https://bugs.webkit.org/show_bug.cgi?id=48366
3366         <rdar://problem/8595202>
3367
3368         * MiniBrowser/mac/BrowserWindowController.m:
3369         (-[BrowserWindowController awakeFromNib]):
3370         * WebKitTestRunner/TestController.cpp:
3371         (WTR::createOtherPage):
3372         (WTR::TestController::initialize):
3373
3374 2010-10-26  Eric Seidel  <eric@webkit.org>
3375
3376         Reviewed by Tony Chang.
3377
3378         Teach webkit-patch how to read credentials from the environment
3379         https://bugs.webkit.org/show_bug.cgi?id=48275
3380
3381         This makes it possible for svn users to have their bugzilla credentials
3382         stored in their environment instead of typing them every time.
3383
3384         We need this for making it easy to run the win-ews bot (which currently
3385         uses svn instead of git).
3386
3387         * Scripts/webkitpy/common/net/credentials.py:
3388         * Scripts/webkitpy/common/net/credentials_unittest.py:
3389
3390 2010-10-26  Kenichi Ishibashi  <bashi@google.com>
3391
3392         Reviewed by Kent Tamura.
3393
3394         Input Method inserts conversion candidates unexpectedly
3395         https://bugs.webkit.org/show_bug.cgi?id=46868
3396
3397         Adds setComposition() to TextInputController to make DRT emulate
3398         an input method behavior.
3399
3400         * DumpRenderTree/chromium/TextInputController.cpp:
3401         (TextInputController::TextInputController):
3402         (TextInputController::setComposition): Added.
3403         * DumpRenderTree/chromium/TextInputController.h:
3404
3405 2010-10-26  Simon Fraser  <simon.fraser@apple.com>
3406
3407         Reviewed by Sam Weinig.
3408
3409         Support layoutTestController.layerTreeAsText in WebKitTestRunner
3410         https://bugs.webkit.org/show_bug.cgi?id=42145
3411         
3412         Implement layerTreeAsText() in WebKitTestRunner.
3413
3414         * WebKitTestRunner/InjectedBundle/Bindings/LayoutTestController.idl:
3415         * WebKitTestRunner/InjectedBundle/LayoutTestController.cpp:
3416         (WTR::LayoutTestController::layerTreeAsText):
3417         * WebKitTestRunner/InjectedBundle/LayoutTestController.h:
3418
3419 2010-10-26  Adam Barth  <abarth@webkit.org>
3420
3421         Reviewed by Eric Seidel.
3422
3423         webkit-patch rollout produces incorrect patch when using svn move
3424         https://bugs.webkit.org/show_bug.cgi?id=48244
3425
3426         We need to flush our caches when we modify the working copy.
3427
3428         * Scripts/webkitpy/tool/steps/revertrevision.py:
3429
3430 2010-10-26  Eric Seidel  <eric@webkit.org>
3431
3432         Reviewed by Adam Barth.
3433
3434         commit-queue flaky test message can list the same author more than once
3435         https://bugs.webkit.org/show_bug.cgi?id=48268
3436
3437         tonikitoo reported to me over IRC this morning that he's seen
3438         the commit-queue report flaky tests with author lists like:
3439         "adam, adam and adam", suggesting we're not uniquing authors
3440         before writing the message.
3441
3442         I fixed the uniquing and added a bunch more unit testing.
3443
3444         * Scripts/webkitpy/tool/commands/queues.py:
3445         * Scripts/webkitpy/tool/commands/queues_unittest.py:
3446         * Scripts/webkitpy/tool/mocktool.py:
3447
3448 2010-10-26  Adam Roben  <aroben@apple.com>
3449
3450         Pull in the FeatureDefines*.vsprops files when building DRT
3451
3452         This ensures that various ENABLE() macros will get set correctly.
3453         (Prior to r70320 we were picking up the ENABLE(3D_RENDERING) flag
3454         through wtf/Platform.h.)
3455
3456         Fixes <http://webkit.org/b/48343> REGRESSION (r70320): DumpRenderTree
3457         says 3D_RENDERING is disabled even when it is enabled
3458
3459         Reviewed by Ada Chan.
3460
3461         * DumpRenderTree/win/DumpRenderTree.vcproj: Added
3462         FeaturesDefines*.vsprops to all configurations.
3463
3464         * DumpRenderTree/win/DumpRenderTreeWin.h: Touched to force a rebuild.
3465
3466 2010-10-26  Søren Gjesse  <sgjesse@chromium.org>
3467
3468         Reviewed by Dimitri Glazkov.
3469
3470         Added options --multiple-loads and --js-flags to chromium DumpRenderTree. The option --multiple-loads=X
3471         is used to have DumpRenderTree load each test it runs X times. To be able to have more fine-grained control
3472         of how the JavaScript engine behaves for each load the flag --js-flags can specify a list of flag-sets like this
3473
3474           --js-flags="--xxx,--noxxx --yyy,--noyyy"
3475
3476         First load will run with --xxx, the second with --yyy and the third without any (the 'no' prefix is handled by
3477         V8 to turn off the flag).
3478
3479         The changes to the Python test runner will be in a separate change.
3480
3481
3482         * DumpRenderTree/chromium/DumpRenderTree.cpp:
3483         (runTest):
3484         (main):
3485         * DumpRenderTree/chromium/TestShell.cpp:
3486         (TestShell::TestShell):
3487         (TestShell::runFileTest):
3488         (TestShell::testFinished):
3489         * DumpRenderTree/chromium/TestShell.h:
3490         (TestShell::loadCount):
3491         (TestShell::setLoadCount):
3492         (TestShell::javaScriptFlagsForLoad):
3493         (TestShell::setJavaScriptFlags):
3494         (TestShell::setDumpWhenFinished):
3495
3496 2010-10-26  Adam Roben  <aroben@apple.com>
3497
3498         Skip more tests that depend on accelerated compositing when accelerated
3499         compositing is disabled
3500
3501         Fixes <http://webkit.org/b/48329> Some tests fail when accelerated
3502         compositing is disabled
3503
3504         Reviewed by John Sullivan.
3505
3506         * Scripts/old-run-webkit-tests: Added more tests to skip when
3507         accelerated compositing is disable. The tests all have output that
3508         changes depending on whether accelerated compositing is enabled.
3509
3510 2010-10-25  Eric Seidel  <eric@webkit.org>
3511
3512         Reviewed by Adam Barth.
3513
3514         webkit-patch should clean up svn locks when passed --force-clean
3515         https://bugs.webkit.org/show_bug.cgi?id=48269
3516
3517         * Scripts/webkitpy/common/checkout/scm.py:
3518
3519 2010-10-25  Anders Carlsson  <andersca@apple.com>
3520
3521         Reviewed by Adam Roben.
3522
3523         Add a pageDidScroll BundleUIClient callback
3524         https://bugs.webkit.org/show_bug.cgi?id=48260
3525         <rdar://problem/8531159>
3526
3527         * WebKitTestRunner/InjectedBundle/InjectedBundlePage.cpp:
3528         (WTR::InjectedBundlePage::InjectedBundlePage):
3529
3530 2010-10-25  Johnny Ding  <jnd@chromium.org>
3531
3532         Reviewed by Tony Chang.
3533
3534         Dump the gesture status of frame in frame load callbacks in DumpRenderTree
3535         by adding a new method dumpUserGestureInFrameLoadCallbacks.
3536         Now only dump the gesture status in "DidStartProvisionalLoad" callback.
3537         https://bugs.webkit.org/show_bug.cgi?id=47849
3538
3539         * DumpRenderTree/LayoutTestController.cpp:
3540         (LayoutTestController::LayoutTestController):
3541         (dumpUserGestureInFrameLoadCallbacksCallback):
3542         (LayoutTestController::staticFunctions):
3543         * DumpRenderTree/LayoutTestController.h:
3544         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
3545         (LayoutTestController::setDumpUserGestureInFrameLoadCallbacks):
3546         * DumpRenderTree/chromium/LayoutTestController.cpp:
3547         (LayoutTestController::LayoutTestController):
3548         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
3549         (LayoutTestController::reset):
3550         * DumpRenderTree/chromium/LayoutTestController.h:
3551         (LayoutTestController::shouldDumpUserGestureInFrameLoadCallbacks):
3552         (LayoutTestController::setShouldDumpUserGestureInFrameLoadCallbacks):
3553         * DumpRenderTree/chromium/TestShell.h:
3554         (TestShell::shouldDumpUserGestureInFrameLoadCallbacks):
3555         * DumpRenderTree/chromium/WebViewHost.cpp:
3556         (WebViewHost::didStartProvisionalLoad):
3557         (WebViewHost::printFrameUserGestureStatus):
3558         * DumpRenderTree/chromium/WebViewHost.h:
3559         * DumpRenderTree/mac/FrameLoadDelegate.mm:
3560         (-[WebFrame _drt_printFrameUserGestureStatus]):
3561         (-[FrameLoadDelegate webView:didStartProvisionalLoadForFrame:]):
3562         (-[FrameLoadDelegate webView:didCommitLoadForFrame:]):
3563         (-[FrameLoadDelegate webView:didFailProvisionalLoadWithError:forFrame:]):
3564         (-[FrameLoadDelegate webView:didFinishLoadForFrame:]):
3565         (-[FrameLoadDelegate webView:didFailLoadWithError:forFrame:]):
3566         (-[FrameLoadDelegate webView:windowScriptObjectAvailable:]):
3567         (-[FrameLoadDelegate webView:didReceiveTitle:forFrame:]):
3568         * DumpRenderTree/qt/LayoutTestControllerQt.cpp:
3569         (LayoutTestController::reset):
3570         (LayoutTestController::dumpUserGestureInFrameLoadCallbacks):
3571         * DumpRenderTree/qt/LayoutTestControllerQt.h: